Home | History | Annotate | Line # | Download | only in notes
testfloat-history.txt revision 1.1.1.1
      1  1.1  ross 
      2  1.1  ross History of Major Changes to TestFloat, up to Release 2a
      3  1.1  ross 
      4  1.1  ross John R. Hauser
      5  1.1  ross 1998 December 17
      6  1.1  ross 
      7  1.1  ross 
      8  1.1  ross The TestFloat releases parallel those of SoftFloat, on which TestFloat is
      9  1.1  ross based.  Each TestFloat release also incorporates all bug fixes from the
     10  1.1  ross corresponding release of SoftFloat.
     11  1.1  ross 
     12  1.1  ross - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
     13  1.1  ross Release 2a (1998 December)
     14  1.1  ross 
     15  1.1  ross -- Added support for testing conversions between floating-point and 64-bit
     16  1.1  ross    integers.
     17  1.1  ross 
     18  1.1  ross -- Improved the makefiles.
     19  1.1  ross 
     20  1.1  ross - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
     21  1.1  ross Release 2 (1997 June)
     22  1.1  ross 
     23  1.1  ross -- Integrated the generation of test cases and the checking of system
     24  1.1  ross    results into a single program.  (Before they were separate programs,
     25  1.1  ross    normally joined by explicit command-line pipes.)
     26  1.1  ross 
     27  1.1  ross -- Improved the sequence of test cases.
     28  1.1  ross 
     29  1.1  ross -- Added support for testing extended double precision and quadruple
     30  1.1  ross    precision.
     31  1.1  ross 
     32  1.1  ross -- Made program output more readable, and added new command arguments.
     33  1.1  ross 
     34  1.1  ross -- Reduced dependence on the quality of the standard `random' function for
     35  1.1  ross    generating test cases.  (Previously naively expected `random' to be able
     36  1.1  ross    to generate good random bits for the entire machine word width.)
     37  1.1  ross 
     38  1.1  ross -- Created `testsoftfloat', with its own simpler complete software floating-
     39  1.1  ross    point (``slowfloat'') for comparison purposes.
     40  1.1  ross 
     41  1.1  ross -- Made some changes to the source file structure, including renaming
     42  1.1  ross    `environment.h' to `milieu.h' (to avoid confusion with environment
     43  1.1  ross    variables).
     44  1.1  ross 
     45  1.1  ross - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
     46  1.1  ross Release 1a (1996 July)
     47  1.1  ross 
     48  1.1  ross -- Added the `-tininessbefore' and `-tininessafter' options to control
     49  1.1  ross    whether tininess should be detected before or after rounding.
     50  1.1  ross 
     51  1.1  ross - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
     52  1.1  ross Release 1 (1996 July)
     53  1.1  ross 
     54  1.1  ross -- Original release.
     55  1.1  ross 
     56  1.1  ross - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
     57  1.1  ross 
     58