AVX2: The BLENDPW instruction selects between vectors of v16i16 using an i8
authorNadav Rotem <nadav.rotem@intel.com>
Tue, 24 Apr 2012 11:27:53 +0000 (11:27 +0000)
committerNadav Rotem <nadav.rotem@intel.com>
Tue, 24 Apr 2012 11:27:53 +0000 (11:27 +0000)
commit87ffdbcb7b93db35d8ff87dfb84d6ae623a5f49f
tree031ea94a62b73318bbcf6ce7616cc5b2809da0bb
parentf4478f99dd63503bf0f0e763bc6d684e738bfe3d
AVX2: The BLENDPW instruction selects between vectors of v16i16 using an i8
immediate. We can't use it here because the shuffle code does not check that
the lower part of the word is identical to the upper part.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155440 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Target/X86/X86ISelLowering.cpp