This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Kahan's Floating Point Test
- To: Michael Price <mprice at atl dot lmco dot com>
- Subject: Re: Kahan's Floating Point Test
- From: "M. Edward (Ed) Borasky" <znmeb at aracnet dot com>
- Date: Wed, 30 May 2001 16:49:40 -0700 (PDT)
- cc: <moshier at moshier dot ne dot mediaone dot net>, <gcc at gcc dot gnu dot org>
On Wed, 30 May 2001, Michael Price wrote:
> On Wed, May 30, 2001 at 01:54:38PM -0400, Stephen L Moshier wrote:
> >
> > See the FAQ at http://gcc.gnu.org, in the section on bugs and
> > non-bugs, for info on floating point behavior.
>
> I've read the FAQ. Perhaps I'm not making my point clear.
>
> You can write assembly on the x86 that produces results that conform to IEEE
> 754. This is not done by default or with any compiler switches on the x86,
> apparently because the conformant code isn't "fast" enough for most peoples
> tastes while the "almost" conformant code is good enough for most people.
>
> My questions is: Why isn't there a switch like
> -mieee754-compliant-code-no-matter-how-slow-it-is-on-this-buggy-x86-fpu
>
> I realize that poorly written libm functions can still cause problems but
> that is relatively easy to fix.
I have Red Hat 7.1 / gcc pre-release 3.0 / Athlon Thunderbird. Is there any
reason I should test this, or do we know it will fail (or pass)??
--
znmeb@aracnet.com (M. Edward Borasky) http://www.aracnet.com/~znmeb
Q. Who invented the non-Von Neumann computer architecture?
A. John non-Von Neumann.