[PATCH] D52059: [InstCombine][x86] try harder to convert blendv intrinsic to generic IR (PR38814)

Fri Sep 14 06:57:29 PDT 2018

spatel updated this revision to Diff 165495.
spatel added a comment.

Patch updated:
No code changes, but added tests based on the ugly bitcasted reality when using SSE intrinsics in C code. The motivating tests from PR38814 also include loads, compares, and constant operands in case we want to restrict this transform by including any of those elements in the pattern matching.

https://reviews.llvm.org/D52059

Files:
  lib/Transforms/InstCombine/InstCombineCalls.cpp
  test/Transforms/InstCombine/X86/blend_x86.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D52059.165495.patch
Type: text/x-patch
Size: 6165 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20180914/c65e07b7/attachment.bin>