This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [patch] Support vectorization of double reduction
- From: Ira Rosen <IRAR at il dot ibm dot com>
- To: Richard Guenther <richard dot guenther at gmail dot com>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Mon, 13 Jul 2009 13:26:14 +0300
- Subject: Re: [patch] Support vectorization of double reduction
Richard Guenther <richard.guenther@gmail.com> wrote on 13/07/2009 11:44:35:
> On Sun, Jul 12, 2009 at 9:10 AM, Ira Rosen<IRAR@il.ibm.com> wrote:
> >
> > Committed revision 149526.
>
> This causes the build of 416.gamess in SPEC 2006 to ICE
> on x86_64 with
>
> /gcc/spec/sb-barbella.suse.de-head-64-2006/x86_64/install-hack/bin/gfortran
> -c -o solib.fppized.o -O3 -funroll-loops -fpeel-loops -ffast-math
> -march=native -DSPEC_CPU_LP64 -ffixed-form
> solib.fppized.f
> solib.fppized.f: In function 'propagate':
> solib.fppized.f:828:0: internal compiler error: in vinfo_for_stmt, at
> tree-vectorizer.h:617
> Please submit a full bug report,
> with preprocessed source if appropriate.
> See <http://gcc.gnu.org/bugs.html> for instructions.
> specmake: *** [solib.fppized.o] Error 1
> specmake: *** Waiting for unfinished jobs....
The bug is in negligent checks in double reduction detection.
I am testing a fix.
Thanks,
Ira
>
> Richard.