diff options
author | bde <bde@FreeBSD.org> | 2006-01-05 09:18:48 +0000 |
---|---|---|
committer | bde <bde@FreeBSD.org> | 2006-01-05 09:18:48 +0000 |
commit | 76448b86540fdcf11b76be6f9cd2a3a1226178f2 (patch) | |
tree | 273f31b0e64905e67f2441c7087d4f837b168a01 /usr.bin/mail | |
parent | eab4be14c141ba8d4dc24c20a0ca8da7c0933a5a (diff) | |
download | FreeBSD-src-76448b86540fdcf11b76be6f9cd2a3a1226178f2.zip FreeBSD-src-76448b86540fdcf11b76be6f9cd2a3a1226178f2.tar.gz |
Oops, on amd64 (and probably on all non-i386 systems), the previous
commit broke the 2**24 cases where |x| > DBL_MAX/2. There are exponent
range problems not just for denormals (underflow) but for large values
(overflow). Doubles have more than enough exponent range to avoid the
problems, but I forgot to convert enough terms to double, so there was
an x+x term which was sometimes evaluated in float precision.
Unfortunately, this is a pessimization with some combinations of systems
and compilers (it makes no difference on Athlon XP's, but on Athlon64's
it gives a 5% pessimization with gcc-3.4 but not with gcc-3.3).
Exlain the problem better in comments.
Diffstat (limited to 'usr.bin/mail')
0 files changed, 0 insertions, 0 deletions