[llvm-commits] [llvm] r166937 - in /llvm/trunk: docs/ReleaseNotes.html include/llvm/Transforms/IPO/PassManagerBuilder.h lib/Transforms/IPO/PassManagerBuilder.cpp

[Renato Golin]

While I see the value in having each vectorization pass individually set,
--vectorize should actually just turn all vectorization passes on, by
default.

--renato


On 29 October 2012 17:35, Hal Finkel <hfinkel at anl.gov> wrote:

> ----- Original Message -----
> > From: "Nadav Rotem" <nrotem at apple.com>
> > To: "Hal Finkel" <hfinkel at anl.gov>
> > Cc: llvm-commits at cs.uiuc.edu
> > Sent: Monday, October 29, 2012 12:17:40 PM
> > Subject: Re: [llvm-commits] [llvm] r166937 - in /llvm/trunk:
> docs/ReleaseNotes.html
> > include/llvm/Transforms/IPO/PassManagerBuilder.h
> lib/Transforms/IPO/PassManagerBuilder.cpp
> >
> >
> >
> >
> > On Oct 29, 2012, at 10:13 AM, Hal Finkel < hfinkel at anl.gov > wrote:
> >
> >
> > I'm okay with this for now, although I'd like to merge them back
> > together at some point.
> >
> >
> >
> > I split the two passes because we need to be able to run one or the
> > other. It is difficult to debug and benchmark when both passes are
> > enabled.
>
> Fair enough, but I'd like to make sure that we keep buildbot coverage on
> the loop vectorizer. That is my primary concern with the current splitting.
> Another option is to keep the current flag and add options to turn off loop
> vectorization or bb-vectorization.
>
>  -Hal
>
> --
> Hal Finkel
> Postdoctoral Appointee
> Leadership Computing Facility
> Argonne National Laboratory
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
>



-- 
cheers,
--renato

http://systemcall.org/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20121101/8fdda54d/attachment.html>