This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: * Re: What is acceptable for -ffast-math? (Was: associative law in combine)
- To: moshier at moshier dot ne dot mediaone dot net, torvalds at transmeta dot com
- Subject: Re: * Re: What is acceptable for -ffast-math? (Was: associative law in combine)
- From: dewar at gnat dot com
- Date: Sun, 29 Jul 2001 17:20:21 -0400 (EDT)
- Cc: gcc at gcc dot gnu dot org, tprince at computer dot org
<<But most people who write floating point code do not think that way. They
consider FP to be "mathematically exact" (which _does_ obey associative
and distributive laws), and they write their programs that way.
>>
I don't know if it is true that most people do not know what they are doing
when they right floating-point code, but that's irrelevant. The default
behavior should be to behave correctly, the fact that many programmers
don't know what correct behavior means is no excuse to do otherwise.
In fact many programmers lack knowledge of many fine points of programming
language definitions, but we do not use that as an excuse for not bothering
to follow the definition.