[PATCH][AVX512] Add 512b integer shift by variable patterns and intrinsics

Cameron McInally cameron.mcinally at nyu.edu
Tue Nov 18 13:32:41 PST 2014


On Tue, Nov 18, 2014 at 3:33 PM, Cameron McInally
<cameron.mcinally at nyu.edu> wrote:
> Hey guys,
>
> Attached is a patch to support 512b integer shift by variable
> intrinsics for AVX512.

Just noticed that the base is incorrect on the multiclasses. It should
be AVX512BIBase, not AVX512BIi8Base.



More information about the llvm-commits mailing list