[PATCH] [X86, AVX] use blends instead of insert128 with index 0

[Sanjay Patel]

Hi andreadb, chandlerc, bruno,

Another case of x86-specific shuffle strength reduction: avoid generating insert*128 instructions with index 0 because they are slower than their non-lane-changing blend equivalents.

http://reviews.llvm.org/D8366

Files:
  lib/Target/X86/X86ISelLowering.cpp
  test/CodeGen/X86/2012-04-26-sdglue.ll
  test/CodeGen/X86/avx-cast.ll

EMAIL PREFERENCES
  http://reviews.llvm.org/settings/panel/emailpreferences/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: D8366.22057.patch
Type: text/x-patch
Size: 6271 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20150316/1d3cf5a1/attachment.bin>