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

[Elena Demikhovsky via llvm-commits]

delena requested changes to this revision.
delena added a comment.
This revision now requires changes to proceed.

Please wait! I disagree with adding a bunch of unmasked intrinsics additionally to the masked. If you want to create IR in InstCombineCalls, you can do this for masked intrinsics as well. I'm afraid that we will need to duplicate hundreds of intrinsics. I want to ask Intel guys opinion before commit.


https://reviews.llvm.org/D26333