[PATCH] D80032: [AMDGPU] Always expand ext/insertelement with divergent idx

[Stanislav Mekhanoshin via Phabricator via llvm-commits]

rampitec created this revision.
rampitec added a reviewer: arsenm.
Herald added subscribers: kerbowa, hiraditya, t-tye, tpr, dstuttard, yaxunl, nhaehnle, wdng, jvesely, kzhuravl.
Herald added a project: LLVM.

Even though series of cmd/cndmask can produce quite a lot of
code that is still better than a loop. In case of doubles we
would even produce two loops.


https://reviews.llvm.org/D80032

Files:
  llvm/lib/Target/AMDGPU/SIISelLowering.cpp
  llvm/test/CodeGen/AMDGPU/extract_vector_dynelt.ll
  llvm/test/CodeGen/AMDGPU/indirect-addressing-si-gfx9.ll
  llvm/test/CodeGen/AMDGPU/indirect-addressing-si-pregfx9.ll
  llvm/test/CodeGen/AMDGPU/indirect-addressing-si.ll
  llvm/test/CodeGen/AMDGPU/insert_vector_dynelt.ll
  llvm/test/CodeGen/AMDGPU/scratch-simple.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D80032.264332.patch
Type: text/x-patch
Size: 18610 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20200515/b0367461/attachment.bin>