This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: tree-ssa performance
- From: Gerald Pfeifer <pfeifer at dbai dot tuwien dot ac dot at>
- To: Jeffrey A Law <law at redhat dot com>
- Cc: Steven Bosscher <s dot bosscher at student dot tudelft dot nl>,Andrew MacLeod <amacleod at redhat dot com>,Diego Novillo <dnovillo at redhat dot com>, Andrew Haley <aph at redhat dot com>,Jeff Sturm <jsturm at one-point dot com>, Tim Josling <tej at melbpc dot org dot au>,Zack Weinberg <zack at codesourcery dot com>, doko at cs dot tu-berlin dot de,gcc mailing list <gcc at gcc dot gnu dot org>, Neil Booth <neil at daikokuya dot co dot uk>
- Date: Wed, 11 Jun 2003 15:30:30 +0200 (CEST)
- Subject: Re: tree-ssa performance
- References: <200306101705.h5AH5AHe023356@speedy.slc.redhat.com>
On Tue, 10 Jun 2003 law@redhat.com wrote:
> Our handling of EH is far from optimal from a compile-time performance
> standpoint due to how we handle cleanups. I've got a patch here which
> addresses that problem, but it's queued behind a couple other things
> I have to nail down.
To give you some motivation to push it to the head of the queue...
> Changing the #if 0 around "skip-cleanups = 0" in tree-cfg.c to an #if 1
> should give folks an idea of whether or not my pending patch will help
> (though don't try to use such a compiler for real work :-) Gerald, if
> you could try that, it would be helpful.
...here you are some results of today's tree-ssa branch without and with
this change for the testcase from PR8361. :-)
-O0
37.25
35.39
-O1
98.32
95.25
-O2
160.24
119.83
-O3
162.72
126.49
(Timings may be a bit inaccurate, but the trend is quite clear, especially
for -O2 and -O3!)
Gerald
PS: I strongly recommend http://gcc.gnu.org/PR8361 for regular use in
assessing your improvements.