This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [BENCHMARK]-mfpmath=sse should disable x387 intrinsics
- From: Jakub Jelinek <jakub at redhat dot com>
- To: Roger Sayle <roger at eyesopen dot com>
- Cc: Richard Guenther <richard dot guenther at gmail dot com>, Uros Bizjak <uros at kss-loka dot si>, gcc-patches at gcc dot gnu dot org
- Date: Thu, 25 Nov 2004 13:07:25 -0500
- Subject: Re: [BENCHMARK]-mfpmath=sse should disable x387 intrinsics
- References: <84fc9c000411250554298d36c1@mail.gmail.com> <Pine.LNX.4.44.0411250934230.12169-100000@www.eyesopen.com>
- Reply-to: Jakub Jelinek <jakub at redhat dot com>
On Thu, Nov 25, 2004 at 09:40:12AM -0700, Roger Sayle wrote:
> You could try -mno-fp-ret-in-387 which produces much prettier code but
> breaks standard ABIs. Indeed, the 386.c backend could probably be tweaked
> to implictly perform the equivalent of "-mno-fp-ret-in-387" when using
> "-mfpmath=sse" for static functions within a single compilation unit.
> Alternatively, it could also perform the equivalent of a hypothetical
> "-mfp-ret-in-sse" in exactly these cases too.
Yeah, for static functions it should probably also pass say float and
double arguments in SSE registers as well, not on the stack, not just
vector arguments.
Jakub