[X86] Custom split v32i16/v64i8 bitcasts when AVX512F is available, but BWI is not.
authorCraig Topper <craig.topper@intel.com>
Fri, 23 Feb 2018 18:43:36 +0000 (18:43 +0000)
committerCraig Topper <craig.topper@intel.com>
Fri, 23 Feb 2018 18:43:36 +0000 (18:43 +0000)
commit11704dcc720e64e473253740bd894adb7f430ab8
tree6849cd1d346813249472160ff6b845db40fd9bd4
parentf8bf2ec0a82ed7c10b64102ea37c004e2865c378
[X86] Custom split v32i16/v64i8 bitcasts when AVX512F is available, but BWI is not.

The test changes you can see are related to the changes in ReplaceNodeResults. Though shuffle-vs-trunc-512.ll does have a test that exercises the code in LowerBITCAST. Looks like the test output didn't change because DAG combining is able to clean up the resulting type legalization. Adding the custom hook just makes type legalization work less hard.

Differential Revision: https://reviews.llvm.org/D43447

llvm-svn: 325933
llvm/lib/Target/X86/X86ISelLowering.cpp
llvm/test/CodeGen/X86/avx512vl-vec-masked-cmp.ll