This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Timings for reassociation pass
- From: Daniel Berlin <dberlin at dberlin dot org>
- To: Richard Guenther <richard dot guenther at gmail dot com>
- Cc: Diego Novillo <dnovillo at redhat dot com>, GCC Patches <gcc-patches at gcc dot gnu dot org>
- Date: Sun, 15 May 2005 11:06:59 -0400
- Subject: Re: Timings for reassociation pass
- References: <1116125094.31071.6.camel@linux.site> <84fc9c0005051504257e3713d1@mail.gmail.com>
On Sun, 2005-05-15 at 13:25 +0200, Richard Guenther wrote:
> On 5/15/05, Daniel Berlin <dberlin@dberlin.org> wrote:
> > For the curious, some timings for the simple reassociation pass at:
> > http://gcc.gnu.org/ml/gcc-patches/2005-05/msg00751.html
> >
> > For tramp3d, out of 386 seconds, reassociation takes .24 seconds (0%).
> > Almost all the time is in hash lookups.
> >
> > By comparison, splitting critical edges takes longer :)
> >
> > tree split crit edges : 0.25 ( 0%) usr
> > tree reassociation : 0.24 ( 0%) usr
> >
> > It takes roughly the same relative % for cc1-i-files, and MICO as well.
>
> I'm curious - does it buy us anything on these benchmarks? I mean
> f.i. in runtime for tramp3d?
Nope. It only finds a few things to reassociate in tramp3d (10 or so),
last i looked.