This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: x86_64 gen_align fix
- From: Jan Hubicka <jh at suse dot cz>
- To: Roger Sayle <roger at eyesopen dot com>
- Cc: Jan Hubicka <jh at suse dot cz>, gcc-patches at gcc dot gnu dot org,Andreas Jaeger <aj at suse dot de>
- Date: Sat, 7 Jun 2003 14:48:02 +0200
- Subject: Re: x86_64 gen_align fix
- References: <Pine.LNX.4.44.0306062053480.15022-100000@www.eyesopen.com>
>
> Hi Jan,
>
> Andreas' SPEC2000 testing seems to have uncovered a dramatic
> performance degradation on mainline CVS originating sometime
> yesterday:
>
> http://www.suse.de/~aj/SPEC/CINT/d-permanent/recent.html
>
> For example, 164.gzip performance drops from 476 to 441 (about -7%).
>
> Examining his log files, the regression occured between the Jun 4,
> 22:18 UTC check-out and the Jun 5, 13:20 UTC check-out. It turns
> out that only a single patch was committed in that period:
>
> > Thu Jun 5 13:23:51 CEST 2003 Jan Hubicka <jh@suse.cz>
> >
> > * i386.md (align): Fix warning; clarify what to do when no
> > p2align is available.
>
>
> Any chance that you could take a look at it?
>
> Many thanks in advance,
Actually it is not this patch, it is the earlier patch for aligning the
jumps but tree was broken for while and Andrea's scripts are not
recording the diffs of failed builds.
the problem is that I by accident do the alignment even for Athln where
it is loss. I will add the missing if (TARGET_K8) soon.
Honza
>
> Roger
> --
>