This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug tree-optimization/54776] [4.8 Regression] tramp3d-v4: 20% performance regression using -O3
- From: "markus at trippelsdorf dot de" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Thu, 04 Oct 2012 19:40:27 +0000
- Subject: [Bug tree-optimization/54776] [4.8 Regression] tramp3d-v4: 20% performance regression using -O3
- Auto-submitted: auto-generated
- References: <bug-54776-4@http.gcc.gnu.org/bugzilla/>
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54776
--- Comment #8 from Markus Trippelsdorf <markus at trippelsdorf dot de> 2012-10-04 19:40:27 UTC ---
(In reply to comment #7)
> > > I assume you do see improvements with -O3 -fwhole-program? (it improves the
> > > scores in my setup, I would be curious to see how it compares to other numbers
> > > you report)
> >
> > Yes, with -O3 -fwhole-program I get:
> > Time spent in iteration: 6.34161
> >
> > (With -fprofile-use -Ofast -march=native:
> > Time spent in iteration: 2.75908 )
>
> Hmm, interesting. Fixing the current resolution bug should get -flto and
> -fwhole-program scores equivalent, but obviously there is a lot room for
> improvement - I did not see so much preivously.
>
> How does text size compare with
> clang lto/non-lto
1003656/1332016
> gcc whole-program/non-whole-program?
1518688/1722480/(1873392(-flto))