[PATCH] [X86] replace vinsertf128 intrinsics with generic shuffles

Sanjay Patel spatel at rotateright.com
Fri Mar 6 10:25:39 PST 2015


Patch updated:

1. Mask off all but the low bit of the immediate
2. Added test case to confirm masking


http://reviews.llvm.org/D8086

Files:
  include/llvm/IR/IntrinsicsX86.td
  lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp
  lib/IR/AutoUpgrade.cpp
  test/CodeGen/X86/avx-intrinsics-x86-upgrade.ll
  test/CodeGen/X86/avx-intrinsics-x86.ll
  test/CodeGen/X86/avx-vinsertf128.ll
  test/CodeGen/X86/unaligned-32-byte-memops.ll

EMAIL PREFERENCES
  http://reviews.llvm.org/settings/panel/emailpreferences/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: D8086.21371.patch
Type: text/x-patch
Size: 16025 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20150306/bf4acc63/attachment.bin>


More information about the llvm-commits mailing list