This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH]: Alignment analysis from autovect
- From: Daniel Berlin <dberlin at dberlin dot org>
- To: Andi Kleen <ak at muc dot de>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Sun, 27 Mar 2005 13:37:25 -0500
- Subject: Re: [PATCH]: Alignment analysis from autovect
- References: <1111696993.9378.8.camel@linux.site> <m1psxl2bmy.fsf@muc.de>
On Sun, 2005-03-27 at 20:21 +0200, Andi Kleen wrote:
> Daniel Berlin <dberlin@dberlin.org> writes:
>
> > As part of the vectorizer improvements, this adds alignment analysis
> > that was on autovect, that uses the propagation engine to attempt to
> > determine pointer alignment.
> >
> > There is room for improvement in what it can handle (for example, it
> > doesn't deal with masking operations right now), but it works, and is
> > useful to the vectorizer.
>
> [...]
>
> Is the pointer alignment generated by this patch used by memset/memcpy
> builtin code generation too?
Not yet.
There are plenty of things that can use the results of this analysis, of
course, and memset/memcpy just one of them.