This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: auto vectorization in gcc
- From: law at redhat dot com
- To: Michael Matz <matz at suse dot de>
- Cc: gcc at gcc dot gnu dot org, Dorit Naishlos <DORIT at il dot ibm dot com>, Diego Novillo <dnovillo at redhat dot com>, David Edelsohn <dje at watson dot ibm dot com>, Richard Henderson <rth at redhat dot com>
- Date: Thu, 17 Jul 2003 12:11:48 -0600
- Subject: Re: auto vectorization in gcc
- Reply-to: law at redhat dot com
In message <Pine.LNX.4.44.0307172000580.10372-100000@wotan.suse.de>, Michael Ma
tz writes:
>Hi,
>
>On Thu, 17 Jul 2003, Richard Henderson wrote:
>
>> The tree level is *more* capable than the rtl level at representing
>> vector types (and thus operations). I think all we need is some small
>> amount of info from the target about vector widths and memory blocking,
>> and then the transformation should happen at the tree level.
>
>I agree. Autovectorization should be much easier on tree level. The
>required transformations are complicated enough as is, without having to
>deal with RTL issues.
What I've always envisioned is that we'd do the vectorization at the tree
level, building vectors as wide as possible (up to some limit). Then
at the tree->rtl phase we'd break the vectors down to whatever size the
target actually supports. The basic idea being to not have a lot of
target dependencies in the vectorizer.
jeff