This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: PR 23551: why should we coalesce inlined variables?
- From: "Andrew Pinski" <pinskia at gmail dot com>
- To: "Alexandre Oliva" <aoliva at redhat dot com>
- Cc: "Andrew MacLeod" <amacleod at redhat dot com>, gcc-patches at gcc dot gnu dot org
- Date: Fri, 1 Jun 2007 17:15:52 -0700
- Subject: Re: PR 23551: why should we coalesce inlined variables?
- References: <or7irhxh3t.fsf@free.oliva.athome.lsd.ic.unicamp.br> <1178807417.3671.249.camel@localhost.localdomain> <orabw67iep.fsf@free.oliva.athome.lsd.ic.unicamp.br> <1179318946.3671.478.camel@localhost.localdomain> <orr6ovgqsy.fsf@free.oliva.athome.lsd.ic.unicamp.br>
On 6/1/07, Alexandre Oliva <aoliva@redhat.com> wrote:
I didn't. Today I set some time aside to try to get SPEC2000 to run
on an x86_64 box with Fedora 7. I still haven't got all of the
benchmarks to compile and run successfully with the GCC trunk, but the
results I got for the patch are promising:
Again, the left column is the run-time WITH the patch, the right
columnt is the run-time WITHOUT the patch. That's right, removing the
patch actually slows things down. I couldn't quite believe it, after
what you said, so I triple-checked.
Oh I can believe it for x86_64 :) You might want to try on more than
x86_64, the register allocator might not be doing a good job for
x86_64. Without the patch is causing more register pressure than with
the patch. Try either on ia64 or PPC where you have more registers.
In that case with the patch might slow down the runtime. If that
happens, then the summary we can come to, is that your patch reduces
register pressure (and maybe causes loop optimizers not to do as much
as they did before likewise for PRE and other interesting optimizers).
Thanks,
Andrew Pinski