This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [ tree-ssa] Eliminate more GOTO_EXPRs
- From: law at redhat dot com
- To: Andrew Pinski <pinskia at physics dot uc dot edu>
- Cc: Jason Merrill <jason at redhat dot com>, gcc-patches at gcc dot gnu dot org
- Date: Tue, 20 May 2003 14:52:23 -0600
- Subject: Re: [ tree-ssa] Eliminate more GOTO_EXPRs
- Reply-to: law at redhat dot com
In message <E0EA039A-8AFC-11D7-AF3E-000393A6D2F2@physics.uc.edu>, Andrew Pinski
writes:
>What about looking at insn-recog.c, which has loads of goto's in it,
>does the tree-ssa branch do better job than the mainline with the hack?
I don't have a copy of the mainline tree handy. I'm doing basically no
work on the mainline these days.
However, I can compare with and without -fdisable-simple, which should
be roughly equivalent to comparing the tree-ssa branch with the mainline
sources.
-O2 -fdisable-simple 119369 insns
-O2 (with hack) 124019 insns
As you can see on that code we're generating 4-5% more insns with the
gimplifier + tree-ssa optimizations.
Some of those are spurious line notes (no clue where they're coming from),
some are from unnecessary labels, and some from poor handling of
partial-word conditionals. At least that's what I saw with a very very
quick glance at the file.
Jeff