[PATCH] D26333: [AVX-512] Add unmasked version of shift by immediate and shift by single element in XMM.

[Simon Pilgrim via llvm-commits]

RKSimon accepted this revision.
RKSimon added a comment.
This revision is now accepted and ready to land.

LGTM

In https://reviews.llvm.org/D26333#591505, @delena wrote:

> Do you mean InstCombine on IR-2-IR or DAG combine?


InstCombineCalls.cpp has code to convert SSE/AVX2 vector shifts to generic shifts if we can guarantee that the shift values are in range. With this change Craig will be able to add support for the AVX512 equivalents without having to add mask/maskz support.


https://reviews.llvm.org/D26333