This is the mail archive of the
mailing list for the GCC project.
Re: Compilation performance comparison of 3.5.0 and TreeSSA trees on MICO sources as requested in: [tree-ssa] Merge status 2004-05-03
- From: Steven Bosscher <stevenb at suse dot de>
- To: Diego Novillo <dnovillo at redhat dot com>
- Cc: Karel Gardas <kgardas at objectsecurity dot com>,"gcc at gcc dot gnu dot org" <gcc at gcc dot gnu dot org>
- Date: Wed, 5 May 2004 00:34:27 +0200 (CEST)
- Subject: Re: Compilation performance comparison of 3.5.0 and TreeSSA trees on MICO sources as requested in: [tree-ssa] Merge status 2004-05-03
- Organization: SuSE Linux AG
- References: <Pine.LNX.firstname.lastname@example.org><email@example.com>
On May 05, 2004 12:10 AM, Diego Novillo <firstname.lastname@example.org> wrote:
> On Tue, 2004-05-04 at 16:50, Karel Gardas wrote:
> > Conclusion: compile time regressions for both -O0 and -O2 compilation.
Thanks for all your timings. Can you please time one more thing: What do thinks
look like for you with the following options:
"-O2 -fno-gcse -fno-delete-null-pointer-checks -fno-cse-follow-jumps"
> Thanks. -O0 is a known loser right now. The gimplification process
> tends to create extremely bloated RTL. That can only be fixed by
> introducing simplistic tree optimization passes at O0. I see this as a
> post-merge exercise, as well.
Some of this is already being worked on in the tree profiling branch (where
we have a tree CFG at -O0 on which we can do simple cleanups and such).