[PATCH] D30660: AMDGPU: Don't wait at end of block with a trivial successor

Mon Mar 6 09:05:01 PST 2017

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

If there is only one successor, and that successor only
has one predecessor the wait can obviously be delayed until
uses or the end of the next block. This avoids code quality
regressions when there are trivial fallthrough blocks inserted
for structurization.

https://reviews.llvm.org/D30660

Files:
  lib/Target/AMDGPU/SIInsertWaits.cpp
  test/CodeGen/AMDGPU/waitcnt.mir

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D30660.90714.patch
Type: text/x-patch
Size: 2946 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20170306/92f9c670/attachment.bin>