[PATCH] D32606: AMDGPU: Try to use op_sel when selecting packed instructions

[Matt Arsenault via Phabricator via llvm-commits]

arsenm created this revision.
Herald added subscribers: t-tye, tpr, dstuttard, yaxunl, nhaehnle, wdng, kzhuravl.

Avoids instructions to pack a vector when the source is really
a scalar being broadcast.

      

Also be smarter and look for per-component fneg.

      

Doesn't yet handle scalar from upper half of register
or other swizzles.


https://reviews.llvm.org/D32606

Files:
  lib/Target/AMDGPU/AMDGPUISelDAGToDAG.cpp
  test/CodeGen/AMDGPU/immv216.ll
  test/CodeGen/AMDGPU/packed-op-sel.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D32606.96953.patch
Type: text/x-patch
Size: 11251 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20170427/dfa14b50/attachment.bin>