| History log of /src/external/lgpl3/mpfr/dist |
| Revision | Date | Author | Comments |
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | branches: 1.1.1.2.24; 1.1.1.2.26;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.2.26.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.24.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.7 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | branches: 1.1.1.6.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.6.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.7 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | branches: 1.1.1.6.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.6.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.7 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | branches: 1.1.1.6.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.6.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.7 | 05-Mar-2023 |
mrg | merge mpfr 4.2.0.
|
| 1.6 | 26-Sep-2020 |
mrg | merge mpfr 4.1.0
|
| 1.5 | 04-Sep-2018 |
mrg | merge mpfr 4.0.1.
|
| 1.4 | 17-Aug-2017 |
mrg | branches: 1.4.2; 1.4.4;
merge MPFR 3.1.5.
|
| 1.3 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.2 | 21-Jun-2011 |
mrg | branches: 1.2.2; 1.2.8;
add some "NetBSD_Disabled*" rules to avoid r/o src tree build issues.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.2.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.2.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.4.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.4.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.7 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | branches: 1.1.1.6.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.6.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.4 | 26-Sep-2020 |
mrg | empty PATCHES is needed. revert a local change that is wrong now.
|
| 1.3 | 26-Sep-2020 |
mrg | we don't have a patch version for a while now.
|
| 1.2 | 06-Jul-2011 |
drochner | update mpfr to PL4 (minor fixes)
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.8 | 01-Jul-2024 |
mrg | merge mpfr 4.2.1.
|
| 1.7 | 05-Mar-2023 |
mrg | branches: 1.7.2;
merge mpfr 4.2.0.
|
| 1.6 | 26-Sep-2020 |
mrg | merge mpfr 4.1.0
|
| 1.5 | 04-Sep-2018 |
mrg | merge mpfr 4.0.1.
|
| 1.4 | 17-Aug-2017 |
mrg | branches: 1.4.2; 1.4.4;
merge MPFR 3.1.5.
|
| 1.3 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.2 | 06-Jul-2011 |
drochner | branches: 1.2.2; 1.2.8;
update mpfr to PL4 (minor fixes)
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.7 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.2.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.2.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.4.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.4.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.7.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.7 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | branches: 1.1.1.6.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.6.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 17-Aug-2017 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.4 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.3 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.2 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1 | 17-Aug-2017 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.3 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.2 | 06-Jul-2011 |
drochner | branches: 1.2.2; 1.2.8;
update mpfr to PL4 (minor fixes)
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.2.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.2.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.3 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.2 | 06-Jul-2011 |
drochner | branches: 1.2.2; 1.2.8;
update mpfr to PL4 (minor fixes)
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.2.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.2.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.8 | 05-Mar-2023 |
mrg | merge mpfr 4.2.0.
|
| 1.7 | 17-Nov-2020 |
rin | Support aarch64eb in */config.guess.
Cherry-picked from upstream:
https://git.savannah.gnu.org/gitweb/?p=config.git;a=commit;h=1c4398015583eb77bc043234f5734be055e64bea
Everything except external/apache2/llvm/dist/llvm/cmake/config.guess
is patched, which is under vendor tag and cannot be modified. I expect
that this file is not actually used as we use hand-crafted version of
configure script instead of cmake for building LLVM.
Note that external/apache2/llvm/autoconf/autoconf/config.guess has
already been committed on Oct. 20, but commit message disappeared as
cvs aborted due to "permission denied" when trying to modify the file
mentioned above. Sorry for confusing you.
Also note that GMP uses its own config.guess Patch for
external/lgpl3/gmp/dist/config.guess is provided by ryo@. Thanks!
|
| 1.6 | 26-Sep-2020 |
mrg | merge mpfr 4.1.0
|
| 1.5 | 04-Sep-2018 |
mrg | merge mpfr 4.0.1.
|
| 1.4 | 01-Feb-2017 |
sevan | branches: 1.4.10; 1.4.12;
Update supporting files for components which rely on autoconf to allow systems
introducing since release of software to be recognised. This should hopefully
allow the builds to progress a littles further on systems such as the POWER8
which features a little endian 64-bit PowerPC CPU identified as ppc64le.
|
| 1.3 | 28-Nov-2013 |
mrg | branches: 1.3.8; 1.3.12;
merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.2 | 19-Jul-2012 |
christos | branches: 1.2.2;
PR/46704: Antti Kantee: Recognize tilera in config.guess
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.2.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.2.1 | 30-Oct-2012 |
yamt | sync with head
|
| 1.2.2.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.3.12.1 | 21-Apr-2017 |
bouyer | Sync with HEAD
|
| 1.3.8.1 | 20-Mar-2017 |
pgoyette | Sync with HEAD
|
| 1.4.12.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.4.10.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.5 | 05-Mar-2023 |
mrg | merge mpfr 4.2.0.
|
| 1.4 | 26-Sep-2020 |
mrg | merge mpfr 4.1.0
|
| 1.3 | 04-Sep-2018 |
mrg | merge mpfr 4.0.1.
|
| 1.2 | 01-Feb-2017 |
sevan | branches: 1.2.10; 1.2.12;
Update supporting files for components which rely on autoconf to allow systems
introducing since release of software to be recognised. This should hopefully
allow the builds to progress a littles further on systems such as the POWER8
which features a little endian 64-bit PowerPC CPU identified as ppc64le.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | branches: 1.1.1.2.8; 1.1.1.2.12;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.2.12.1 | 21-Apr-2017 |
bouyer | Sync with HEAD
|
| 1.1.1.2.8.1 | 20-Mar-2017 |
pgoyette | Sync with HEAD
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2.12.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.2.10.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.7 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | branches: 1.1.1.6.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.6.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file configure.ac was added on branch tls-maxphys on 2014-08-20 00:00:03 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file configure.ac was added on branch yamt-pagecache on 2014-05-22 14:09:15 +0000
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.3 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.2 | 06-Jul-2011 |
drochner | branches: 1.2.2; 1.2.8;
update mpfr to PL4 (minor fixes)
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.2.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.2.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file mpfr.pc.in was added on branch phil-wifi on 2019-06-10 22:02:26 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file mpfr.pc.in was added on branch pgoyette-compat on 2018-09-06 06:53:46 +0000
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.3 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.2 | 06-Jul-2011 |
drochner | branches: 1.2.2; 1.2.8;
update mpfr to PL4 (minor fixes)
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.2.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.2.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 17-Aug-2017 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 17-Aug-2017 |
mrg | initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.3 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.2 | 06-Jul-2011 |
drochner | branches: 1.2.2; 1.2.8;
update mpfr to PL4 (minor fixes)
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.2.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.2.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file FAQ.html was added on branch tls-maxphys on 2014-08-20 00:00:04 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file FAQ.html was added on branch yamt-pagecache on 2014-05-22 14:09:15 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file Makefile.am was added on branch tls-maxphys on 2014-08-20 00:00:04 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file Makefile.am was added on branch yamt-pagecache on 2014-05-22 14:09:15 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file Makefile.in was added on branch tls-maxphys on 2014-08-20 00:00:04 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file Makefile.in was added on branch yamt-pagecache on 2014-05-22 14:09:15 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.4 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | branches: 1.1.1.3.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file README.dev was added on branch phil-wifi on 2019-06-10 22:02:26 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file README.dev was added on branch pgoyette-compat on 2018-09-06 06:53:46 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file check-typography was added on branch phil-wifi on 2019-06-10 22:02:27 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file check-typography was added on branch pgoyette-compat on 2018-09-06 06:53:46 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file fdl.texi was added on branch tls-maxphys on 2014-08-20 00:00:04 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file fdl.texi was added on branch yamt-pagecache on 2014-05-22 14:09:15 +0000
|
| 1.1 | 26-Sep-2020 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file mpfr.info was added on branch tls-maxphys on 2014-08-20 00:00:04 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file mpfr.info was added on branch yamt-pagecache on 2014-05-22 14:09:15 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file mpfr.texi was added on branch tls-maxphys on 2014-08-20 00:00:04 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file mpfr.texi was added on branch yamt-pagecache on 2014-05-22 14:09:15 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file texinfo.tex was added on branch tls-maxphys on 2014-08-20 00:00:04 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file texinfo.tex was added on branch yamt-pagecache on 2014-05-22 14:09:15 +0000
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file can_round.c was added on branch phil-wifi on 2019-06-10 22:02:27 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file can_round.c was added on branch pgoyette-compat on 2018-09-06 06:53:46 +0000
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 26-Sep-2020 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.4 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.4 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.4 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2 | 04-Sep-2018 |
mrg | merge mpfr 4.0.1.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file Makefile.am was added on branch tls-maxphys on 2014-08-20 00:00:04 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file Makefile.am was added on branch yamt-pagecache on 2014-05-22 14:09:15 +0000
|
| 1.8 | 01-Jul-2024 |
mrg | merge mpfr 4.2.1.
|
| 1.7 | 05-Mar-2023 |
mrg | branches: 1.7.2;
merge mpfr 4.2.0.
|
| 1.6 | 26-Sep-2020 |
mrg | empty PATCHES is needed. revert a local change that is wrong now.
|
| 1.5 | 26-Sep-2020 |
mrg | merge mpfr 4.1.0
|
| 1.4 | 04-Sep-2018 |
mrg | merge mpfr 4.0.1.
|
| 1.3 | 17-Aug-2017 |
mrg | branches: 1.3.2; 1.3.4;
merge MPFR 3.1.5.
|
| 1.2 | 28-Nov-2013 |
mrg | branches: 1.2.4; 1.2.8;
merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.2.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.2.8.1 | 28-Nov-2013 |
tls | file Makefile.in was added on branch tls-maxphys on 2014-08-20 00:00:04 +0000
|
| 1.2.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2.4.1 | 28-Nov-2013 |
yamt | file Makefile.in was added on branch yamt-pagecache on 2014-05-22 14:09:15 +0000
|
| 1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.7.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file abort_prec_max.c was added on branch tls-maxphys on 2014-08-20 00:00:04 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file abort_prec_max.c was added on branch yamt-pagecache on 2014-05-22 14:09:15 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file acos.c was added on branch tls-maxphys on 2014-08-20 00:00:04 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file acos.c was added on branch yamt-pagecache on 2014-05-22 14:09:15 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file acosh.c was added on branch tls-maxphys on 2014-08-20 00:00:04 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file acosh.c was added on branch yamt-pagecache on 2014-05-22 14:09:15 +0000
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1.1.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file add.c was added on branch tls-maxphys on 2014-08-20 00:00:04 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file add.c was added on branch yamt-pagecache on 2014-05-22 14:09:15 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file add1.c was added on branch tls-maxphys on 2014-08-20 00:00:04 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file add1.c was added on branch yamt-pagecache on 2014-05-22 14:09:15 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file add1sp.c was added on branch tls-maxphys on 2014-08-20 00:00:04 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file add1sp.c was added on branch yamt-pagecache on 2014-05-22 14:09:15 +0000
|
| 1.1 | 26-Sep-2020 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file add_d.c was added on branch tls-maxphys on 2014-08-20 00:00:04 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file add_d.c was added on branch yamt-pagecache on 2014-05-22 14:09:15 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file add_ui.c was added on branch tls-maxphys on 2014-08-20 00:00:04 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file add_ui.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file agm.c was added on branch tls-maxphys on 2014-08-20 00:00:04 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file agm.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file ai.c was added on branch tls-maxphys on 2014-08-20 00:00:04 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file ai.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file asin.c was added on branch tls-maxphys on 2014-08-20 00:00:04 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file asin.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file asinh.c was added on branch tls-maxphys on 2014-08-20 00:00:04 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file asinh.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1.1.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file atan.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file atan.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file atan2.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file atan2.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1.1.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file atanh.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file atanh.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1.1.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file bernoulli.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file bernoulli.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file beta.c was added on branch phil-wifi on 2019-06-10 22:02:27 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file beta.c was added on branch pgoyette-compat on 2018-09-06 06:53:46 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file buildopt.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file buildopt.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file cache.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file cache.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file cbrt.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file cbrt.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file check.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file check.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file clear.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file clear.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file clears.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file clears.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file cmp.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file cmp.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file cmp2.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file cmp2.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.2 | 26-Sep-2020 |
mrg | merge mpfr 4.1.0
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file cmp_abs.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file cmp_abs.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file cmp_d.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file cmp_d.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file cmp_ld.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file cmp_ld.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file cmp_si.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file cmp_si.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file cmp_ui.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file cmp_ui.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 26-Sep-2020 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1 | 26-Sep-2020 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file comparisons.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file comparisons.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1.1.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file const_catalan.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file const_catalan.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file const_euler.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file const_euler.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file const_log2.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file const_log2.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file const_pi.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file const_pi.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file constant.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file constant.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file copysign.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file copysign.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file cos.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file cos.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file cosh.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file cosh.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1.1.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file cot.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file cot.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file coth.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file coth.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file csc.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file csc.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file csch.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file csch.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file d_div.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file d_div.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file d_sub.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file d_sub.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file digamma.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file digamma.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file dim.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file dim.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file div.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file div.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file div_2exp.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file div_2exp.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file div_2si.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file div_2si.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file div_2ui.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file div_2ui.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file div_d.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file div_d.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file div_ui.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file div_ui.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 26-Sep-2020 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file dump.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file dump.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file eint.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file eint.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file eq.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file eq.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file erandom.c was added on branch phil-wifi on 2019-06-10 22:02:27 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file erandom.c was added on branch pgoyette-compat on 2018-09-06 06:53:46 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file erf.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file erf.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file erfc.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file erfc.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file exceptions.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file exceptions.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file exp.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file exp.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file exp10.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file exp10.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1.1.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file exp2.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file exp2.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1.1.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file exp3.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file exp3.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file exp_2.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file exp_2.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file expm1.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file expm1.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file extract.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file extract.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file factorial.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file factorial.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file fits_intmax.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file fits_intmax.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file fits_s.h was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file fits_s.h was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file fits_sint.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file fits_sint.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file fits_slong.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file fits_slong.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file fits_sshort.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file fits_sshort.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file fits_u.h was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file fits_u.h was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file fits_uint.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file fits_uint.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file fits_uintmax.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file fits_uintmax.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file fits_ulong.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file fits_ulong.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file fits_ushort.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file fits_ushort.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file fma.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file fma.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.4 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | branches: 1.1.1.3.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file fmma.c was added on branch phil-wifi on 2019-06-10 22:02:27 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file fmma.c was added on branch pgoyette-compat on 2018-09-06 06:53:46 +0000
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1.1.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file fms.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file fms.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file fpif.c was added on branch phil-wifi on 2019-06-10 22:02:27 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file fpif.c was added on branch pgoyette-compat on 2018-09-06 06:53:46 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file frac.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file frac.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file free_cache.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file free_cache.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file frexp.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file frexp.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file gamma.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file gamma.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file gamma_inc.c was added on branch phil-wifi on 2019-06-10 22:02:27 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file gamma_inc.c was added on branch pgoyette-compat on 2018-09-06 06:53:46 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file gammaonethird.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file gammaonethird.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file gen_inverse.h was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file gen_inverse.h was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file get_d.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file get_d.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 26-Sep-2020 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file get_d64.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file get_d64.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file get_exp.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file get_exp.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file get_f.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file get_f.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file get_float128.c was added on branch phil-wifi on 2019-06-10 22:02:27 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file get_float128.c was added on branch pgoyette-compat on 2018-09-06 06:53:46 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file get_flt.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file get_flt.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file get_ld.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file get_ld.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.2 | 26-Sep-2020 |
mrg | merge mpfr 4.1.0
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file get_patches.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file get_patches.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file get_q.c was added on branch phil-wifi on 2019-06-10 22:02:27 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file get_q.c was added on branch pgoyette-compat on 2018-09-06 06:53:46 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file get_si.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file get_si.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file get_sj.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file get_sj.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file get_str.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file get_str.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file get_ui.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file get_ui.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file get_uj.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file get_uj.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file get_z.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file get_z.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.2 | 05-Mar-2023 |
mrg | merge mpfr 4.2.0.
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file get_z_exp.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file get_z_exp.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file gmp_op.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file gmp_op.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file grandom.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file grandom.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file hypot.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file hypot.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file ieee_floats.h was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file ieee_floats.h was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file init.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file init.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file init2.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file init2.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file inits.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file inits.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file inits2.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file inits2.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file inp_str.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file inp_str.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file int_ceil_log2.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file int_ceil_log2.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file invert_limb.h was added on branch phil-wifi on 2019-06-10 22:02:27 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file invert_limb.h was added on branch pgoyette-compat on 2018-09-06 06:53:46 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file invsqrt_limb.h was added on branch phil-wifi on 2019-06-10 22:02:28 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file invsqrt_limb.h was added on branch pgoyette-compat on 2018-09-06 06:53:46 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file isinf.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file isinf.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file isinteger.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file isinteger.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file isnan.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file isnan.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file isnum.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file isnum.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file isqrt.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file isqrt.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file isregular.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file isregular.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file iszero.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file iszero.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file jn.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file jn.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file jyn_asympt.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file jyn_asympt.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file li2.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file li2.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file lngamma.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file lngamma.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file log.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file log.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file log10.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file log10.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1.1.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file log1p.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file log1p.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file log2.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file log2.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1.1.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file log_ui.c was added on branch phil-wifi on 2019-06-10 22:02:28 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file log_ui.c was added on branch pgoyette-compat on 2018-09-06 06:53:46 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file logging.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file logging.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file min_prec.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file min_prec.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file minmax.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file minmax.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file modf.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file modf.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file mp_clz_tab.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file mp_clz_tab.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file mparam_h.in was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file mparam_h.in was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file mpf2mpfr.h was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file mpf2mpfr.h was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file mpfr-cvers.h was added on branch phil-wifi on 2019-06-10 22:02:28 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file mpfr-cvers.h was added on branch pgoyette-compat on 2018-09-06 06:53:46 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file mpfr-gmp.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file mpfr-gmp.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file mpfr-gmp.h was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file mpfr-gmp.h was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file mpfr-impl.h was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file mpfr-impl.h was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file mpfr-intmax.h was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file mpfr-intmax.h was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.6 | 05-Mar-2023 |
mrg | merge mpfr 4.2.0.
|
| 1.5 | 26-Sep-2020 |
mrg | merge mpfr 4.1.0
|
| 1.4 | 04-Sep-2018 |
mrg | merge mpfr 4.0.1.
|
| 1.3 | 17-Aug-2017 |
mrg | branches: 1.3.2; 1.3.4;
merge MPFR 3.1.5.
|
| 1.2 | 25-Mar-2014 |
joerg | branches: 1.2.4; 1.2.8;
Use C version of umul_ppmm for Clang on MIPS.
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.2.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.2.8.1 | 25-Mar-2014 |
tls | file mpfr-longlong.h was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.2.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.2.4.1 | 25-Mar-2014 |
yamt | file mpfr-longlong.h was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file mpfr-mini-gmp.c was added on branch phil-wifi on 2019-06-10 22:02:28 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file mpfr-mini-gmp.c was added on branch pgoyette-compat on 2018-09-06 06:53:46 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file mpfr-mini-gmp.h was added on branch phil-wifi on 2019-06-10 22:02:28 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file mpfr-mini-gmp.h was added on branch pgoyette-compat on 2018-09-06 06:53:46 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file mpfr-sassert.h was added on branch phil-wifi on 2019-06-10 22:02:28 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file mpfr-sassert.h was added on branch pgoyette-compat on 2018-09-06 06:53:46 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file mpfr-thread.h was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file mpfr-thread.h was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file mpfr.h was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file mpfr.h was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file mpn_exp.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file mpn_exp.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file mul.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file mul.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 26-Sep-2020 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file mul_2exp.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file mul_2exp.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file mul_2si.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file mul_2si.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file mul_2ui.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file mul_2ui.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file mul_d.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file mul_d.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file mul_ui.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file mul_ui.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file mulders.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file mulders.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 26-Sep-2020 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file neg.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file neg.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file next.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file next.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file nrandom.c was added on branch phil-wifi on 2019-06-10 22:02:28 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file nrandom.c was added on branch pgoyette-compat on 2018-09-06 06:53:47 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file odd_p.c was added on branch phil-wifi on 2019-06-10 22:02:28 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file odd_p.c was added on branch pgoyette-compat on 2018-09-06 06:53:47 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file out_str.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file out_str.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file pool.c was added on branch phil-wifi on 2019-06-10 22:02:28 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file pool.c was added on branch pgoyette-compat on 2018-09-06 06:53:47 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file pow.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file pow.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file pow_si.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file pow_si.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file pow_ui.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file pow_ui.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file pow_z.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file pow_z.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file powerof2.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file powerof2.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file print_raw.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file print_raw.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file print_rnd_mode.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file print_rnd_mode.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file printf.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file printf.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file random_deviate.c was added on branch phil-wifi on 2019-06-10 22:02:28 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file random_deviate.c was added on branch pgoyette-compat on 2018-09-06 06:53:47 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file random_deviate.h was added on branch phil-wifi on 2019-06-10 22:02:28 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file random_deviate.h was added on branch pgoyette-compat on 2018-09-06 06:53:47 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file rec_sqrt.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file rec_sqrt.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file reldiff.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file reldiff.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file rem1.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file rem1.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file rint.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file rint.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file rndna.c was added on branch phil-wifi on 2019-06-10 22:02:28 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file rndna.c was added on branch pgoyette-compat on 2018-09-06 06:53:47 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file root.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file root.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file round_near_x.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file round_near_x.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file round_p.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file round_p.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file round_prec.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file round_prec.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file round_raw_generic.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file round_raw_generic.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file scale2.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file scale2.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file sec.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file sec.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file sech.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file sech.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file set.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file set.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file set_d.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file set_d.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 26-Sep-2020 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file set_d64.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file set_d64.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file set_dfl_prec.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file set_dfl_prec.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file set_exp.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file set_exp.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file set_f.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file set_f.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file set_float128.c was added on branch phil-wifi on 2019-06-10 22:02:28 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file set_float128.c was added on branch pgoyette-compat on 2018-09-06 06:53:47 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file set_flt.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file set_flt.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file set_inf.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file set_inf.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file set_ld.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file set_ld.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file set_nan.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file set_nan.c was added on branch yamt-pagecache on 2014-05-22 14:09:16 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file set_prc_raw.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file set_prc_raw.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file set_prec.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file set_prec.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file set_q.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file set_q.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file set_rnd.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file set_rnd.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file set_si.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file set_si.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file set_si_2exp.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file set_si_2exp.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file set_sj.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file set_sj.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file set_str.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file set_str.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file set_str_raw.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file set_str_raw.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file set_ui.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file set_ui.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file set_ui_2exp.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file set_ui_2exp.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file set_uj.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file set_uj.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file set_z.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file set_z.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.2 | 05-Mar-2023 |
mrg | merge mpfr 4.2.0.
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file set_z_exp.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file set_z_exp.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file set_zero.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file set_zero.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file setmax.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file setmax.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file setmin.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file setmin.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file setsign.c was added on branch tls-maxphys on 2014-08-20 00:00:05 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file setsign.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file sgn.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file sgn.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file si_op.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file si_op.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file signbit.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file signbit.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file sin.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file sin.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file sin_cos.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file sin_cos.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file sinh.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file sinh.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file sinh_cosh.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file sinh_cosh.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1.1.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file sqr.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file sqr.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file sqrt.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file sqrt.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file sqrt_ui.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file sqrt_ui.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file stack_interface.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file stack_interface.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file strtofr.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file strtofr.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file sub.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file sub.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file sub1.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file sub1.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file sub1sp.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file sub1sp.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 26-Sep-2020 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file sub_d.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file sub_d.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file sub_ui.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file sub_ui.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file subnormal.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file subnormal.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file sum.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file sum.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file swap.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file swap.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file tan.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file tan.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file tanh.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file tanh.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1.1.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1 | 26-Sep-2020 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.4 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | branches: 1.1.1.3.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file ubf.c was added on branch phil-wifi on 2019-06-10 22:02:28 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file ubf.c was added on branch pgoyette-compat on 2018-09-06 06:53:47 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file uceil_exp2.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file uceil_exp2.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file uceil_log2.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file uceil_log2.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file ufloor_log2.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file ufloor_log2.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file ui_div.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file ui_div.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file ui_pow.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file ui_pow.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file ui_pow_ui.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file ui_pow_ui.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file ui_sub.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file ui_sub.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file urandom.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file urandom.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file urandomb.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file urandomb.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.7 | 01-Jul-2024 |
mrg | merge mpfr 4.2.1.
|
| 1.6 | 05-Mar-2023 |
mrg | branches: 1.6.2;
merge mpfr 4.2.0.
|
| 1.5 | 26-Sep-2020 |
mrg | merge mpfr 4.1.0
|
| 1.4 | 04-Sep-2018 |
mrg | merge mpfr 4.0.1.
|
| 1.3 | 17-Aug-2017 |
mrg | branches: 1.3.2; 1.3.4;
merge MPFR 3.1.5.
|
| 1.2 | 27-Jun-2016 |
martin | Apply upstream r8705: fix logic error.
Reported by Henning Petersen in PR toolchain/51283.
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file vasprintf.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file vasprintf.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.6.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file version.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file version.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file volatile.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file volatile.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file yn.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file yn.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file zeta.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file zeta.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | branches: 1.1.1.5.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.5.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file zeta_ui.c was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file zeta_ui.c was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 26-Sep-2020 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file mparam.h was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file mparam.h was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file mparam.h was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file mparam.h was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file mparam.h was added on branch phil-wifi on 2019-06-10 22:02:29 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file mparam.h was added on branch pgoyette-compat on 2018-09-06 06:53:47 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file mparam.h was added on branch phil-wifi on 2019-06-10 22:02:29 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file mparam.h was added on branch pgoyette-compat on 2018-09-06 06:53:48 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file mparam.h was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file mparam.h was added on branch yamt-pagecache on 2014-05-22 14:09:17 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file mparam.h was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file mparam.h was added on branch yamt-pagecache on 2014-05-22 14:09:18 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file mparam.h was added on branch tls-maxphys on 2014-08-20 00:00:06 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file mparam.h was added on branch yamt-pagecache on 2014-05-22 14:09:18 +0000
|
| 1.1 | 26-Sep-2020 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file mparam.h was added on branch tls-maxphys on 2014-08-20 00:00:07 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file mparam.h was added on branch yamt-pagecache on 2014-05-22 14:09:18 +0000
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.7 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | branches: 1.1.1.6.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.6.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.7 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | branches: 1.1.1.6.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.6.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4; 1.1.1.1.16;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.16.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file inp_str.dat was added on branch phil-wifi on 2019-06-10 22:02:30 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file inp_str.dat was added on branch pgoyette-compat on 2018-09-06 06:53:48 +0000
|
| 1.2 | 04-Sep-2018 |
mrg | merge mpfr 4.0.1.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.40; 1.1.1.1.42;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.1.42.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.40.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.7 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | branches: 1.1.1.6.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.6.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file tabort_defalloc1.c was added on branch phil-wifi on 2019-06-10 22:02:31 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file tabort_defalloc1.c was added on branch pgoyette-compat on 2018-09-06 06:53:48 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file tabort_defalloc2.c was added on branch phil-wifi on 2019-06-10 22:02:31 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file tabort_defalloc2.c was added on branch pgoyette-compat on 2018-09-06 06:53:48 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file tabort_prec_max.c was added on branch phil-wifi on 2019-06-10 22:02:31 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file tabort_prec_max.c was added on branch pgoyette-compat on 2018-09-06 06:53:48 +0000
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file talloc-cache.c was added on branch phil-wifi on 2019-06-10 22:02:31 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file talloc-cache.c was added on branch pgoyette-compat on 2018-09-06 06:53:48 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file talloc.c was added on branch phil-wifi on 2019-06-10 22:02:31 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file talloc.c was added on branch pgoyette-compat on 2018-09-06 06:53:48 +0000
|
| 1.7 | 05-Mar-2023 |
mrg | merge mpfr 4.2.0.
|
| 1.6 | 26-Sep-2020 |
mrg | merge mpfr 4.1.0
|
| 1.5 | 04-Sep-2018 |
mrg | merge mpfr 4.0.1.
|
| 1.4 | 17-Aug-2017 |
mrg | branches: 1.4.2; 1.4.4;
merge MPFR 3.1.5.
|
| 1.3 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.2 | 06-Jul-2011 |
drochner | branches: 1.2.2; 1.2.8;
update mpfr to PL4 (minor fixes)
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.2.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.2.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.4.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.4.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file tassert.c was added on branch phil-wifi on 2019-06-10 22:02:31 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file tassert.c was added on branch pgoyette-compat on 2018-09-06 06:53:48 +0000
|
| 1.7 | 05-Mar-2023 |
mrg | merge mpfr 4.2.0.
|
| 1.6 | 26-Sep-2020 |
mrg | merge mpfr 4.1.0
|
| 1.5 | 04-Sep-2018 |
mrg | merge mpfr 4.0.1.
|
| 1.4 | 17-Aug-2017 |
mrg | branches: 1.4.2; 1.4.4;
merge MPFR 3.1.5.
|
| 1.3 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.2 | 06-Jul-2011 |
drochner | branches: 1.2.2; 1.2.8;
update mpfr to PL4 (minor fixes)
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.2.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.2.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.4.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.4.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file tbeta.c was added on branch phil-wifi on 2019-06-10 22:02:31 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file tbeta.c was added on branch pgoyette-compat on 2018-09-06 06:53:48 +0000
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1.1.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 26-Sep-2020 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file terandom.c was added on branch phil-wifi on 2019-06-10 22:02:31 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file terandom.c was added on branch pgoyette-compat on 2018-09-06 06:53:49 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file terandom_chisq.c was added on branch phil-wifi on 2019-06-10 22:02:31 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file terandom_chisq.c was added on branch pgoyette-compat on 2018-09-06 06:53:49 +0000
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.7 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | branches: 1.1.1.6.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.6.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.7 | 05-Mar-2023 |
mrg | merge mpfr 4.2.0.
|
| 1.6 | 26-Sep-2020 |
mrg | merge mpfr 4.1.0
|
| 1.5 | 04-Sep-2018 |
mrg | merge mpfr 4.0.1.
|
| 1.4 | 17-Aug-2017 |
mrg | branches: 1.4.2; 1.4.4;
merge MPFR 3.1.5.
|
| 1.3 | 28-Nov-2013 |
mrg | merge MPFR 3.1.2. changes to moved srcs are already merged.
|
| 1.2 | 06-Jul-2011 |
drochner | branches: 1.2.2; 1.2.8;
update mpfr to PL4 (minor fixes)
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.2.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.2.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.4.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.4.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.7 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | branches: 1.1.1.6.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.6.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.7 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | branches: 1.1.1.6.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.6.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file tfmma.c was added on branch phil-wifi on 2019-06-10 22:02:31 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file tfmma.c was added on branch pgoyette-compat on 2018-09-06 06:53:49 +0000
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.4 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | branches: 1.1.1.3.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file tfpif.c was added on branch phil-wifi on 2019-06-10 22:02:31 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file tfpif.c was added on branch pgoyette-compat on 2018-09-06 06:53:49 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file tfpif_r1.dat was added on branch phil-wifi on 2019-06-10 22:02:31 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file tfpif_r1.dat was added on branch pgoyette-compat on 2018-09-06 06:53:49 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file tfpif_r2.dat was added on branch phil-wifi on 2019-06-10 22:02:31 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file tfpif_r2.dat was added on branch pgoyette-compat on 2018-09-06 06:53:49 +0000
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.7 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | branches: 1.1.1.6.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.6.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file tfrexp.c was added on branch tls-maxphys on 2014-08-20 00:00:07 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file tfrexp.c was added on branch yamt-pagecache on 2014-05-22 14:09:18 +0000
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file tgamma_inc.c was added on branch phil-wifi on 2019-06-10 22:02:31 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file tgamma_inc.c was added on branch pgoyette-compat on 2018-09-06 06:53:49 +0000
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file tget_q.c was added on branch phil-wifi on 2019-06-10 22:02:31 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file tget_q.c was added on branch pgoyette-compat on 2018-09-06 06:53:49 +0000
|
| 1.1 | 26-Sep-2020 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file tgrandom.c was added on branch tls-maxphys on 2014-08-20 00:00:07 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file tgrandom.c was added on branch yamt-pagecache on 2014-05-22 14:09:18 +0000
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.7 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | branches: 1.1.1.6.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.6.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.7 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | branches: 1.1.1.6.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.6.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.7 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | branches: 1.1.1.6.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.6.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.7 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | branches: 1.1.1.6.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.6.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file tlog_ui.c was added on branch phil-wifi on 2019-06-10 22:02:31 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file tlog_ui.c was added on branch pgoyette-compat on 2018-09-06 06:53:49 +0000
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file tnrandom.c was added on branch phil-wifi on 2019-06-10 22:02:31 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file tnrandom.c was added on branch pgoyette-compat on 2018-09-06 06:53:49 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file tnrandom_chisq.c was added on branch phil-wifi on 2019-06-10 22:02:31 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file tnrandom_chisq.c was added on branch pgoyette-compat on 2018-09-06 06:53:49 +0000
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1 | 26-Sep-2020 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.7 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | branches: 1.1.1.6.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.6.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file trandom_deviate.c was added on branch phil-wifi on 2019-06-10 22:02:31 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file trandom_deviate.c was added on branch pgoyette-compat on 2018-09-06 06:53:49 +0000
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.7 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | branches: 1.1.1.6.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.6.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 01-Jul-2024 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 01-Jul-2024 |
mrg | branches: 1.1.1.1.4;
import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.1.4.2 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.1.4.1 | 01-Jul-2024 |
perseant | file treldiff.c was added on branch perseant-exfatfs on 2025-08-02 05:50:28 +0000
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.7 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | branches: 1.1.1.6.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.6.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file trndna.c was added on branch phil-wifi on 2019-06-10 22:02:31 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file trndna.c was added on branch pgoyette-compat on 2018-09-06 06:53:49 +0000
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file trootn_ui.c was added on branch phil-wifi on 2019-06-10 22:02:31 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file trootn_ui.c was added on branch pgoyette-compat on 2018-09-06 06:53:49 +0000
|
| 1.2 | 26-Sep-2020 |
mrg | merge mpfr 4.1.0
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.4 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | branches: 1.1.1.3.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file tset_float128.c was added on branch phil-wifi on 2019-06-10 22:02:32 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file tset_float128.c was added on branch pgoyette-compat on 2018-09-06 06:53:49 +0000
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.2 | 05-Mar-2023 |
mrg | merge mpfr 4.2.0.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.7 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | branches: 1.1.1.6.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.6.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file tstdint.c was added on branch tls-maxphys on 2014-08-20 00:00:07 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file tstdint.c was added on branch yamt-pagecache on 2014-05-22 14:09:18 +0000
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.7 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | branches: 1.1.1.6.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.6.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 05-Mar-2023 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.1 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1 | 26-Sep-2020 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.7 | 01-Jul-2024 |
mrg | import MPFR 4.2.1.
mostly a bug-fix release, highlights include:
- abort on lock failure, instead of just warn
- better Inf handling
- fix an unlikely stack overflow in mpfr_rec_sqrt()
- fixes for mpfr_reldiff()
- fix boundary error in mpfr_pow_general()
- fixes to printing Nan and Inf
- many manual and test updates
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | branches: 1.1.1.6.2;
initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.6.2.1 | 02-Aug-2025 |
perseant | Sync with HEAD
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file tvalist.c was added on branch tls-maxphys on 2014-08-20 00:00:07 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file tvalist.c was added on branch yamt-pagecache on 2014-05-22 14:09:18 +0000
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.6 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.5 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.4 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.3 | 17-Aug-2017 |
mrg | branches: 1.1.1.3.2; 1.1.1.3.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.2 | 28-Nov-2013 |
mrg | initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.1 | 20-Jun-2011 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.8;
initial import of MPRF 3.0.1.
The MPFR library is a C library for multiple-precision floating-point
computations with exact rounding (also called correct rounding). It is
based on the GMP multiple-precision library and should replace the MPF
class in further releases of GMP.
GCC >= 4.2 requires MPFR.
|
| 1.1.1.3.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.3.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.1 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.2.1 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1 | 26-Sep-2020 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file ck-clz_tab was added on branch phil-wifi on 2019-06-10 22:02:32 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file ck-clz_tab was added on branch pgoyette-compat on 2018-09-06 06:53:49 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file ck-copyright-notice was added on branch tls-maxphys on 2014-08-20 00:00:07 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file ck-copyright-notice was added on branch yamt-pagecache on 2014-05-22 14:09:18 +0000
|
| 1.1 | 26-Sep-2020 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1 | 17-Aug-2017 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.4 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.3 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.2 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1 | 17-Aug-2017 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file ck-version-info was added on branch tls-maxphys on 2014-08-20 00:00:07 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file ck-version-info was added on branch yamt-pagecache on 2014-05-22 14:09:19 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file get_patches.sh was added on branch tls-maxphys on 2014-08-20 00:00:07 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file get_patches.sh was added on branch yamt-pagecache on 2014-05-22 14:09:19 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file Makefile.am was added on branch phil-wifi on 2019-06-10 22:02:32 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file Makefile.am was added on branch pgoyette-compat on 2018-09-06 06:53:49 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file Makefile.in was added on branch phil-wifi on 2019-06-10 22:02:32 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file Makefile.in was added on branch pgoyette-compat on 2018-09-06 06:53:49 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.2 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file README was added on branch phil-wifi on 2019-06-10 22:02:32 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file README was added on branch pgoyette-compat on 2018-09-06 06:53:49 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file benchtime.h was added on branch phil-wifi on 2019-06-10 22:02:32 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file benchtime.h was added on branch pgoyette-compat on 2018-09-06 06:53:49 +0000
|
| 1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.3 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.2 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.1 | 04-Sep-2018 |
mrg | branches: 1.1.1.1.2; 1.1.1.1.4;
import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.1.4.2 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.1.4.1 | 04-Sep-2018 |
christos | file mpfrbench.c was added on branch phil-wifi on 2019-06-10 22:02:32 +0000
|
| 1.1.1.1.2.2 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.2.1 | 04-Sep-2018 |
pgoyette | file mpfrbench.c was added on branch pgoyette-compat on 2018-09-06 06:53:49 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file Makefile.am was added on branch tls-maxphys on 2014-08-20 00:00:07 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file Makefile.am was added on branch yamt-pagecache on 2014-05-22 14:09:19 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file Makefile.in was added on branch tls-maxphys on 2014-08-20 00:00:07 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file Makefile.in was added on branch yamt-pagecache on 2014-05-22 14:09:19 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file bidimensional_sample.c was added on branch tls-maxphys on 2014-08-20 00:00:07 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file bidimensional_sample.c was added on branch yamt-pagecache on 2014-05-22 14:09:19 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file speed.c was added on branch tls-maxphys on 2014-08-20 00:00:07 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file speed.c was added on branch yamt-pagecache on 2014-05-22 14:09:19 +0000
|
| 1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1;
Initial revision
|
| 1.1.1.5 | 05-Mar-2023 |
mrg | initial import of MPFR 4.2.0. changes from 4.1.0 include:
Binary compatible with MPFR 4.0.* and 4.1.*, though some minor changes
in the behavior of the formatted output functions may be visible,
regarded as underspecified behavior or bug fixes (see below).
New functions mpfr_cosu, mpfr_sinu, mpfr_tanu, mpfr_acosu, mpfr_asinu,
mpfr_atanu and mpfr_atan2u.
New functions mpfr_cospi, mpfr_sinpi, mpfr_tanpi, mpfr_acospi,
mpfr_asinpi, mpfr_atanpi and mpfr_atan2pi.
New functions mpfr_log2p1, mpfr_log10p1, mpfr_exp2m1, mpfr_exp10m1 and
mpfr_compound_si.
New functions mpfr_fmod_ui, mpfr_powr, mpfr_pown, mpfr_pow_uj,
mpfr_pow_sj and mpfr_rootn_si (mpfr_pown is actually a macro defined as
an alias for mpfr_pow_sj).
Bug fixes.
- In particular, for the formatted output functions (mpfr_printf, etc.),
the case where the precision consists only of a period has been fixed to
be like .0 as specified in the ISO C standard, and the manual has been
corrected and clarified.
- The macros of the custom interface have also been fixed:
they now behave like functions (except a minor limitation for
mpfr_custom_init_set).
|
| 1.1.1.4 | 26-Sep-2020 |
mrg | GNU mpfr 4.1.0. main changes from 4.0:
Changed __float128 to the type _Float128 specified in ISO/IEC TS 18661.
__float128 is used as a fallback if _Float128 is not supported.
New function mpfr_get_str_ndigits about conversion to a string of digits.
New function mpfr_dot for the dot product (incomplete, experimental).
New functions mpfr_get_decimal128 and mpfr_set_decimal128 (available
only when MPFR has been built with decimal float support).
New function mpfr_cmpabs_ui.
New function mpfr_total_order_p for the IEEE 754 totalOrder predicate.
The mpfr_out_str function now accepts bases from -2 to -36, in order to
follow mpfr_get_str and GMP's mpf_out_str functions (these cases gave an
assertion failure, as with other invalid bases).
Shared caches: cleanup; really detect lock failures (abort in this case).
Improved mpfr_add and mpfr_sub when all operands have a precision
equal to twice the number of bits per word, e.g., 128 bits on a 64-bit
platform.
Optimized the tuning parameters for various architectures.
|
| 1.1.1.3 | 04-Sep-2018 |
mrg | import mpfr 4.0.1. main changes since 3.1.5 are:
Changes from version 4.0.0 to version 4.0.1:
- Bug fixes (see ChangeLog file), in particular in mpfr_div_ui, which
could yield an incorrectly rounded result to nearest when using
different precisions; this bug had been present since the introduction
of mpfr_div_ui, and in MPFR 4.0.0, it was affecting mpfr_div too.
Changes from versions 3.1.* to version 4.0.0:
- Partial support of MPFR_RNDF (faithful rounding).
- New functions: mpfr_fpif_export and mpfr_fpif_import to export and import
numbers in a floating-point interchange format, independent both on the
number of bits per word and on the endianness.
- New function mpfr_fmodquo to return the low bits of the quotient
corresponding to mpfr_fmod.
- New functions mpfr_flags_clear, mpfr_flags_set, mpfr_flags_test,
mpfr_flags_save and mpfr_flags_restore to operate on groups of flags.
- New functions mpfr_set_float128 and mpfr_get_float128 to convert from/to
the __float128 type (requires --enable-float128 and compiler support).
- New functions mpfr_buildopt_float128_p and mpfr_buildopt_sharedcache_p.
- New functions mpfr_rint_roundeven and mpfr_roundeven, completing the
other similar round-to-integer functions for rounding to nearest with
the even-rounding rule.
- New macro mpfr_round_nearest_away to add partial emulation of the
rounding to nearest-away (as defined in IEEE 754-2008).
- New functions mpfr_nrandom and mpfr_erandom to generate random numbers
following normal and exponential distributions respectively.
- New functions mpfr_fmma and mpfr_fmms to compute a*b+c*d and a*b-c*d.
- New function mpfr_rootn_ui, similar to mpfr_root, but agreeing with the
rootn function of the IEEE 754-2008 standard.
- New functions mpfr_log_ui to compute the logarithm of an integer,
mpfr_gamma_inc for the incomplete Gamma function.
- New function mpfr_beta for the Beta function (incomplete, experimental).
- New function mpfr_get_q to convert a floating-point number into rational.
- Dropped K&R C compatibility.
- Major speedup in mpfr_add, mpfr_sub, mpfr_mul, mpfr_div and mpfr_sqrt when
all operands have the same precision and this precision is less than twice
the number of bits per word, e.g., less than 128 on a 64-bit computer.
- Speedup by a factor of almost 2 in the double <--> mpfr conversions
(mpfr_set_d and mpfr_get_d).
- Speedup in mpfr_log1p and mpfr_atanh for small arguments.
- Speedup in the mpfr_const_euler function (contributed by Fredrik Johansson),
in the computation of Bernoulli numbers (used in mpfr_gamma, mpfr_li2,
mpfr_digamma, mpfr_lngamma and mpfr_lgamma), in mpfr_div, in mpfr_fma
and mpfr_fms.
|
| 1.1.1.2 | 17-Aug-2017 |
mrg | branches: 1.1.1.2.2; 1.1.1.2.4;
initial import of MPFR 3.1.5 package. changes since 3.1.2:
Changes from version 3.1.4 to version 3.1.5:
- C++11 compatibility.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.4/#fixed> and ChangeLog file).
- More tests.
Changes from version 3.1.3 to version 3.1.4:
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.3/#fixed> and ChangeLog file).
- MinGW (MS Windows): Added support for thread-safe DLL (shared library).
Changes from version 3.1.2 to version 3.1.3:
- Better support for Automake 1.13+ (now used to generate the tarball).
- Improved MPFR manual.
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.2/#fixed> and ChangeLog file).
|
| 1.1.1.1 | 28-Nov-2013 |
mrg | branches: 1.1.1.1.4; 1.1.1.1.8;
initial import of MPFR 3.1.2. changes since 3.0.1:
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.1/#fixed> or ChangeLog file).
- Bug fixes (see <http://www.mpfr.org/mpfr-3.1.0/#fixed> or ChangeLog file).
- TLS support is now detected automatically. If TLS is supported, MPFR is
built as thread safe by default. To disable TLS explicitly, configure
MPFR with --disable-thread-safe.
- The mpfr_urandom and mpfr_urandomb functions now return identical values
on processors with different word size (assuming the same random seed, and
since the GMP random generator does not depend itself on the word size,
cf http://gmplib.org/list-archives/gmp-devel/2010-September/001642.html).
- The mpfr_add_one_ulp and mpfr_sub_one_ulp macros (which are obsolete and
no more documented) will be removed in a future release.
- Speed improvement for the mpfr_sqr and mpfr_div functions using Mulders'
algorithm. As a consequence, other functions using those routines are
also faster.
- Much faster formatted output (mpfr_printf, etc.) with %Rg and similar.
- New functions mpfr_buildopt_gmpinternals_p, mpfr_buildopt_tune_case,
mpfr_frexp, mpfr_grandom and mpfr_z_sub.
- New divide-by-zero exception (flag) and associated functions.
- Internal change: the logging mechanism has been improved.
- Bug fixes, in particular a huge inefficiency in mpfr_exp (when the
target precision is less than MPFR_EXP_THRESHOLD) on hard-to-round
cases, which can take several minutes.
|
| 1.1.1.2.4.1 | 10-Jun-2019 |
christos | Sync with HEAD
|
| 1.1.1.2.2.1 | 06-Sep-2018 |
pgoyette | Sync with HEAD
Resolve a couple of conflicts (result of the uimin/uimax changes)
|
| 1.1.1.1.8.2 | 20-Aug-2014 |
tls | Rebase to HEAD as of a few days ago.
|
| 1.1.1.1.8.1 | 28-Nov-2013 |
tls | file tuneup.c was added on branch tls-maxphys on 2014-08-20 00:00:07 +0000
|
| 1.1.1.1.4.2 | 22-May-2014 |
yamt | sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
|
| 1.1.1.1.4.1 | 28-Nov-2013 |
yamt | file tuneup.c was added on branch yamt-pagecache on 2014-05-22 14:09:19 +0000
|
