[PATCH] D70304: [Loop Peeling] Add possibility to enable peeling on loop nests.

[Arkady Shlykov via Phabricator via llvm-commits]

ashlykov updated this revision to Diff 230665.
ashlykov added a comment.

@fhahn thanks for your comments, addressed!

Regarding the impact - it highly depends on further simplifications that can be applied to a code after the loop peeling (branch folding, better dependency analysis, reducing of register pressure inside the loop). And since the max number of iterations to be peeled is restricted - the codesize impact is manageable.

As a good case I can point to the nested loops in ImageMagick project (line 3346, https://github.com/ImageMagick/ImageMagick6/blob/master/magick/accelerate-kernels-private.h), I observed a significant gain on the kernel here.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D70304/new/

https://reviews.llvm.org/D70304

Files:
  llvm/include/llvm/Analysis/TargetTransformInfo.h
  llvm/lib/Transforms/Scalar/LoopUnrollPass.cpp
  llvm/lib/Transforms/Utils/LoopUnrollPeel.cpp
  llvm/test/Transforms/LoopUnroll/peel-loop-conditions.ll
  llvm/test/Transforms/LoopUnroll/peel-loop-nests.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D70304.230665.patch
Type: text/x-patch
Size: 17298 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20191125/ef49bb36/attachment-0001.bin>