[PATCH][X86] Teach the backend how to lower vector shift left into multiply rather than scalarizing it.

[Andrea Di Biagio]

Thank you very much! :-)

I committed the patch that introduces the new lowering rules for
vector shifts at revision 201271.
The change to the vectorizer has been commited instead at revision 201272.



On Wed, Feb 12, 2014 at 11:22 PM, Jim Grosbach <grosbach at apple.com> wrote:
> Agreed. This is excellent.
>
> -Jim
>
> On Feb 12, 2014, at 3:20 PM, Nadav Rotem <nrotem at apple.com> wrote:
>
>> LGTM!
>>
>> On Feb 12, 2014, at 1:52 PM, Andrea Di Biagio <andrea.dibiagio at gmail.com> wrote:
>>
>>> <patch-cost-model-v3.diff>
>>
>