This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: tree-ssa memory and compile-time regressions
- From: Gerald Pfeifer <gp at suse dot de>
- To: Richard Guenther <rguenth at tat dot physik dot uni-tuebingen dot de>,Diego Novillo <dnovillo at redhat dot com>
- Cc: gcc at gcc dot gnu dot org, Jan Hubicka <jh at suse dot cz>
- Date: Thu, 8 Jan 2004 17:08:07 +0100 (CET)
- Subject: Re: tree-ssa memory and compile-time regressions
- References: <Pine.LNX.4.53.0401081510250.4680@bellatrix.tat.physik.uni-tuebingen.de>
On Thu, 8 Jan 2004, Richard Guenther wrote:
> Looking at the testcase (stripped down application)
> http://www.tat.physik.uni-tuebingen.de/~rguenth/gcc/apressure.ii.gz
>
> timings are (1GHz Athlon, 1G ram), all with -fno-exceptions:
>
> -O0 220MB, 0m17s
> -O1 260MB, 0m24s
> -O2 800MB, 7m35s (-ftime-report on a faster machine attached below)
> -O2 -fno-unit-at-a-time 260MB, 0m28s
>
> with current HEAD I get
>
> -O0 180MB, 0m15s
> -O1 200MB, 0m19s
> -O2 260MB, 0m28s
This mimics my experience with recent tree-ssa: Despite significant
improvements, I simply cannot build the application behind PR8361 with
-O3 any longer (while that used to work on the very same machine).
Both -fno-unit-at-a-time and...
Diego Novillo <dnovillo@redhat.com> wrote:
> Hmm, based on what the time report says, could you try with
> -fno-tree-pre?
...-fno-tree-pre make the problem go away.
Gerald