This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: patch for merging graphite branch (before tuplification)
- From: Jack Howarth <howarth at bromo dot msbb dot uc dot edu>
- To: Tobias Grosser <grosser at fim dot uni-passau dot de>
- Cc: Sebastian Pop <sebpop at gmail dot com>, Richard Guenther <richard dot guenther at gmail dot com>, GCC Patches <gcc-patches at gcc dot gnu dot org>, Mark Mitchell <mark at codesourcery dot com>, Jakub Jelinek <jakub at redhat dot com>, David Edelsohn <edelsohn at gmail dot com>, "Harle, Christophe" <christophe dot harle at amd dot com>, Konrad Trifunovic <konrad dot trifunovic at gmail dot com>, Albert Cohen <Albert dot Cohen at inria dot fr>, Roberto Bagnara <bagnara at cs dot unipr dot it>
- Date: Thu, 7 Aug 2008 16:09:37 -0400
- Subject: Re: patch for merging graphite branch (before tuplification)
- References: <cb9d34b20807251914jb7fb76q4452be18461d7464@mail.gmail.com> <84fc9c000807260228h12552595x17b2a7556d35913b@mail.gmail.com> <cb9d34b20808021726w3dcb5015o9b256ef393dba02c@mail.gmail.com> <20080807152533.GA13457@bromo.msbb.uc.edu> <1218126793.1346.29.camel@tobilaptop> <20080807183218.GA15799@bromo.msbb.uc.edu> <1218136429.1346.40.camel@tobilaptop>
Tobi,
I'll do a fresh bootstrap in the next day or so with
the current gcc trunk and the proposed patch using ppl 0.9
and will post complete testsuite results. A couple questions.
How well integrated is graphite into gcc and gfortran? My understanding
was that the vectorization in gcc 4.3 was limited due to problems
with the flat array representations used by gfortran causing the
stride information to be opaque to the vectorization optimization
code. Is that addressed in the graphite implementation or does
the vectorization code in gcc trunk need to become graphite saavy?
Jack
ps Also, have the graphite maintainers decided if it will be enabled
by default for -O3. If the vectorization code can leverage graphite,
it would certainly seem like a win-win to enable it for -O3.
On Thu, Aug 07, 2008 at 04:13:49PM -0300, Tobias Grosser wrote:
> >
> > Roberto Bagnara's cursory examination of those results was that
> > these were flaws in the testsuite itself and not real errors.
> > I don't seem to see them in ppl cvs.
>
> Ok. Great to hear.
>
> Can you check, if graphite works even with 0.9 and my patch?
>
> > Jack
> > ps Could you post the diff for the patch you applied to the
> > graphite branch and I'll try that tonight.
>
> ïAttached: graphite-Fix_testcases_for_32_bit_2008-08-07.patch
> Fix testcases for 32 bit
>
> 2008-08-06 Tobias Grosser <grosser@fim.uni-passau.de>
>
>
> diff --git a/gcc/testsuite/gcc.dg/graphite/scop-16.c b/gcc/testsuite/gcc.dg/graphite/scop-16.c
> index 8bb7c1d..b63f238 100644
> --- a/gcc/testsuite/gcc.dg/graphite/scop-16.c
> +++ b/gcc/testsuite/gcc.dg/graphite/scop-16.c
> @@ -1,6 +1,6 @@
> /* { dg-do compile } */
> /* { dg-options "-O2 -floop-block -fdump-tree-graphite-all" } */
> -#define N 100000
> +#define N 10000
> void foo (int);
> int test ()
> {
> diff --git a/gcc/testsuite/gcc.dg/graphite/scop-17.c b/gcc/testsuite/gcc.dg/graphite/scop-17.c
> index 50f38f4..be1ccf2 100644
> --- a/gcc/testsuite/gcc.dg/graphite/scop-17.c
> +++ b/gcc/testsuite/gcc.dg/graphite/scop-17.c
> @@ -1,6 +1,6 @@
> /* { dg-do compile } */
> /* { dg-options "-O2 -floop-block -fdump-tree-graphite-all" } */
> -#define N 100000
> +#define N 10000
> void foo (int);
> int test ()
> {