AVX-512: Fixed a bug in shuffle for v64i8 type

Operation SCALAR_TO_VECTOR for v64i8 and v32i16 should be lowered if BW feature is "on".

Differential Revision: http://reviews.llvm.org/D17994



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@263097 91177308-0d34-0410-b5e6-96231b3b80d8
2 files changed