[PATCH] New early loop unrolling pass
Richard Guenther
richard.guenther@gmail.com
Sun Apr 27 17:05:00 GMT 2008
On Sun, Apr 27, 2008 at 2:13 PM, Dominique Dhumieres <dominiq@lps.ens.fr> wrote:
> Richard,
>
>
> > I re-bootstrapped and tested the following which I'll apply tomorrow
> > if there are no objections.
>
> I have applied the new patch and it still yields worse execution time
> for the optimized variants of induct.f90.
Of course, the patch didn't change. I don't consider this a
show-stopper as it obviously
just exposes bugs in the vectorizer or its cost model. There is
plenty of time to address
this during stage1/2 or ignore it. It would be helpful if you could
provide a reduced
runtime testcase with just one loop that shows this regression.
Thanks,
Richard.
More information about the Gcc-patches
mailing list