Lines Matching refs:WILL
5471 #--WHEN REDUCEX IS USED, THE CODE WILL INEVITABLY BE SLOW.
5509 #--integer quotient will be stored in N
5546 #--WILL BE TOO INEFFICIENT. THE WAY AROUND IT IS THAT
5547 #--(SIGN(INARG)*2**63 + FP2) - SIGN(INARG)*2**63 WILL GIVE
5902 #--WHEN REDUCEX IS USED, THE CODE WILL INEVITABLY BE SLOW.
5940 #--integer quotient will be stored in N
5977 #--WILL BE TOO INEFFICIENT. THE WAY AROUND IT IS THAT
5978 #--(SIGN(INARG)*2**63 + FP2) - SIGN(INARG)*2**63 WILL GIVE
6277 #--ATAN(X) DIRECTLY WILL NEED TO USE A RATIONAL APPROXIMATION
6279 #--WILL INVOLVE A VERY LONG POLYNOMIAL.
6716 # Notes: This will always generate one exception -- inexact. #
6791 # This bound will be used in Step 4. #
6816 # T-2, and T-8 will all be exact --- a property that will #
6827 # |M| <= 16380, and Scale = 2^M. Moreover, exp(X) will #
6833 # approximately M. Thus 6.2 will never cause #
6844 # Notes: For non-zero X, the inexact exception will always be #
6849 # the precaution will be useful in the library version of #
6851 # inputs will be done away with.) #
6872 # Notes: Exp(X) will surely overflow or underflow, depending on #
6884 # Notes: This will return X with the appropriate rounding #
6944 # T-2, and T-8 will all be exact --- a property that will #
7006 # Notes: 10.2 will always create an inexact and return -1 + tiny #
9278 # multiply with a denorm which will cause an unimplemented data type
9279 # exception to be put into the machine which will be caught and corrected
9301 # the src will force the dst to a DENORM value or worse. so, let's
9302 # create an fp multiply that will create the result.
10929 cmp.w %d0, %d1 # will denorm push exp < 0?
Indexes created Wed Oct 29 07:09:46 GMT 2025