[PATCH] D25165: [x86] allow 128/256-bit lowering for copysign vector intrinsics (PR30433)

[Sanjay Patel via llvm-commits]

spatel updated this revision to Diff 73294.
spatel added a comment.

Patch updated:
Simon added cost model and SLP vectorizer tests that are affected by the new legalization parameters. I scanned those diffs, and they all look like improvements, but some adjustments may be warranted to fine-tune the cost model.


https://reviews.llvm.org/D25165

Files:
  lib/Target/X86/X86ISelLowering.cpp
  test/Analysis/CostModel/X86/arith-fp.ll
  test/CodeGen/X86/vec-copysign.ll
  test/Transforms/SLPVectorizer/X86/fcopysign.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D25165.73294.patch
Type: text/x-patch
Size: 61800 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20161003/48b9baa4/attachment.bin>