This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: PING for fwprop merge
- From: "Steven Bosscher" <stevenb dot gcc at gmail dot com>
- To: "Paolo Bonzini" <paolo dot bonzini at lu dot unisi dot ch>
- Cc: "GCC Patches" <gcc-patches at gcc dot gnu dot org>
- Date: Sat, 3 Jun 2006 12:30:41 +0200
- Subject: Re: PING for fwprop merge
- References: <44816095.6010303@lu.unisi.ch> <571f6b510606030314g7b21f702h2c9b6bd64cd14417@mail.gmail.com> <44816192.7050804@lu.unisi.ch>
On 6/3/06, Paolo Bonzini <paolo.bonzini@lu.unisi.ch> wrote:
Steven Bosscher wrote:
> It appears that on ppc-darwin we introduce significant SPEC regressions.
Which tests? I have ppc-darwin and SPEC, even though this will go to 4.3.
As I said, I don't have actually seen this myself. SPEC for
x86/x86_64 still doesn't show any problem, and David E.'s ppc testing
also didn't show any problems. I'm guessing for Darwin maybe we see
some effect that fwprop misses something to handle PIC well, but I'm
not sure and I haven't had the time yet to find out for sure.
As I told you off-list, I don't really have time to figure this out.
Also, I'm trying to re-work CSE using Vlad's patch (thanks Bernd) to
make it work on extended basic blocks. I think fwprop + cse on ebbs
will be a better replacement than fwprop with cse path following
ripped out.
If you still have some time to test on darwin and figure out what is
happening, then that would be great. Otherwise I guess we should
postpone fwprop to gcc 4.3.
Gr.
Steven