[PATCH] D103362: [LoopUnroll] Separate peeling from unrolling

[Nikita Popov via Phabricator via llvm-commits]

nikic updated this revision to Diff 349952.
nikic added a comment.
This revision is now accepted and ready to land.

If an explicit PeelCount has been provided for testing, use it directly and don't try to go through other unroll heuristics. Otherwise we might try to use the explicit test PeelCount together with a full unroll heuristic. (Once again, this doesn't affect the optimization pipeline, just testing.)

Also preserve existing tests by removing explicit unroll counts from tests that already have an explicit peel count. The tests are still reasonable apart from that.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D103362/new/

https://reviews.llvm.org/D103362

Files:
  llvm/include/llvm/Transforms/Utils/UnrollLoop.h
  llvm/lib/Transforms/Scalar/LoopUnrollPass.cpp
  llvm/lib/Transforms/Utils/LoopUnroll.cpp
  llvm/lib/Transforms/Utils/LoopUnrollRuntime.cpp
  llvm/test/Transforms/LoopUnroll/peel-loop-and-unroll.ll
  llvm/test/Transforms/LoopUnroll/pr33437.ll
  llvm/test/Transforms/LoopUnroll/pr45939-peel-count-and-complete-unroll.ll
  llvm/test/Transforms/LoopUnroll/wrong_assert_in_peeling.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D103362.349952.patch
Type: text/x-patch
Size: 26149 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20210604/1ea082ac/attachment.bin>