This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: Vectorization of strided access and bootstrap failure problem
- From: Ira Rosen <IRAR at il dot ibm dot com>
- To: "François-Xavier Coudert" <fxcoudert at gmail dot com>
- Cc: "fortran at gcc dot gnu dot org List" <fortran at gcc dot gnu dot org>, "gcc-patches patches" <gcc-patches at gcc dot gnu dot org>
- Date: Mon, 20 Nov 2006 13:22:28 +0200
- Subject: Re: Vectorization of strided access and bootstrap failure problem
"François-Xavier Coudert" <fxcoudert@gmail.com> wrote on 20/11/2006
13:21:10:
> > However, with current mainline, and with this patch bootstrap *with
> > vectorization enabled* fails on x86. It seems to me that this is the
same
> > failure we've had for a while on mainline on ppc-linux, as reported in
> > pr28752, which has nothing to do with this patch (probably just exposed
by
> > it).
> >
> > I'm thinking, if no objection (within, say, 48 hours), to commit the
patch
> > despite the above, while I continue to try to track down the cause of
the
> > bootstrap failure with vectorization enabled.
>
> Did your regular bootstrap (i.e. without vectorization) include
> Fortran?
Yes, it did.
Ira
> If not, then I would like to suggest you bootstrap with
> Fortran enabled, because the Fortran runtime library (libgfortran)
> compiles a few files with vectorization. We've had a few regressions
> due to that already, and we could do without another.
>
> Thanks,
> FX