[PATCH] Adding the loop vectorizer to the LTO pipeline

[Chandler Carruth]

I'm generally fine with this.

On Fri, Feb 21, 2014 at 2:30 PM, Arnold Schwaighofer <
aschwaighofer at apple.com> wrote:

> On Feb 21, 2014, at 2:15 PM, Rafael Espíndola <rafael.espindola at gmail.com>
> wrote:
>
> > It is probably a good idea.
> >
> > Out of curiosity, have you benchmarked also disabling the vectorizer
> > during "clang -flto" and leaving all the vectorization work for LTO?
> > If not, is that something you plan on doing in the future?
> >
>
> I have not benchmarked this and had no immediate plans to do so. I agree
> this is interesting but I expect results to be a lot noisier.


I'm planning to investigate this as part of a complete overhaul of our LTO
pass pipelines once I finish my current work on the pass manager. There are
*numerous* things that are interesting in the same way here, not just the
vectorizer.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20140221/bea177e2/attachment.html>