This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: 3.0-pre vs 3.0 on Haney speed
- To: Mark Mitchell <mark at codesourcery dot com>
- Subject: Re: 3.0-pre vs 3.0 on Haney speed
- From: Paolo Carlini <pcarlini at unitus dot it>
- Date: Tue, 26 Jun 2001 19:54:05 +0200
- CC: gcc at gcc dot gnu dot org
- References: <29380000.993575768@warlock.codesourcery.com>
Hi,
Mark Mitchell wrote:
> > Could it be related to this patch?
> >
> > 2001-06-09 Mark Mitchell <mark@codesourcery.com>
> >
> > * toplev.c (rest_of_compilation): Remove dead code before
> > purge_addressof.
> >
>
> That patch was only in the tree for a few brief days: it was a huge win,
> except that it didn't work. So, if someone benchmarked during the time
> it was in, that would explain seeing better results than in the actual
> 3.0. Also, we turned off sibling call optimizations for C++ in the
> final weeks before 3.0.
Thanks for your authoritative feedback. May I point you to my follow-up
message:
http://gcc.gnu.org/ml/gcc/2001-06/msg01660.html ???
In the light of those tests, I strongly believe that the specific slow down
in the Complex matrix C++ test is *not* attributable to your patch being
reverted (too bad, by the way...).
Thanks,
Paolo Carlini.