[PATCH] D43327: [X86] Promote 16-bit cmovs to 32-bits

[Craig Topper via Phabricator via llvm-commits]

craig.topper updated this revision to Diff 134781.
craig.topper added a comment.

Just create the cmovs as i32 when they are created. Add a DAG combine for sign_extend_reg to avoid regressing the tests in cmov_promotion.ll

My previous approach was trying to follow closer to how and/or/xor/etc. are promoted, but I don't know that we need to follow that. This patch gets rid of the 16-bit cmovs and doesn't regress any of our tests. That's probably a good enough starting point.


https://reviews.llvm.org/D43327

Files:
  lib/Target/X86/X86ISelLowering.cpp
  test/CodeGen/X86/avx512-insert-extract.ll
  test/CodeGen/X86/avx512-mask-op.ll
  test/CodeGen/X86/avx512-schedule.ll
  test/CodeGen/X86/bool-simplify.ll
  test/CodeGen/X86/select.ll
  test/CodeGen/X86/setcc-lowering.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D43327.134781.patch
Type: text/x-patch
Size: 10239 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20180217/cc9d1329/attachment.bin>