auto vectorization in gcc

Steven Bosscher s.bosscher@student.tudelft.nl
Thu Jul 17 20:30:00 GMT 2003


Op do 17-07-2003, om 21:42 schreef law@redhat.com:
> In message <1058450952.3307.60.camel@frodo.toronto.redhat.com>, Diego Novillo w
> rites:
>  >I don't think there's much risk in duplicating effort at the moment. 
>  >I'm not aware of anybody working on vectorization today.
> Agreed.  Hell, we really don't have anyone even working on the loop
> optimizer for trees yet.  Given that we're going to need an unroller
> and basic loop opts, that might be a great place to start.

Huh? Aldy Hernandez sort-of announced he was going to work on an
autovectorization pass following the Larsen paper
(http://gcc.gnu.org/ml/gcc/2003-06/msg00175.html,
http://citeseer.nj.nec.com/larsen00exploiting.html, etc.).

And dberlin had his loop normalization pass, but this work was put on
hold pending the IV work Sebastian Pop was (is?) doing (Pop how's that
coming? ;-). 

Gr.
Steven



More information about the Gcc mailing list