[PATCH] D139788: [LV] Sink scalar operands and merge regions repeatedly.

[Florian Hahn via Phabricator via llvm-commits]

fhahn updated this revision to Diff 482414.
fhahn added a comment.



In D139788#3987248 <https://reviews.llvm.org/D139788#3987248>, @Ayal wrote:

> Nice catch, ship it!
>
> Would be nice to add a test that exercises three iterations of this while loop...

While adding such a test I found a case where left-over empty blocks between regions prevent further merging. I put up D139927 <https://reviews.llvm.org/D139927> to clean up redundant blocks and updated this patch to use it as well to simplify a case that needs 3 iterations.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D139788/new/

https://reviews.llvm.org/D139788

Files:
  llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
  llvm/lib/Transforms/Vectorize/VPlanTransforms.cpp
  llvm/test/Transforms/LoopVectorize/first-order-recurrence-sink-replicate-region.ll
  llvm/test/Transforms/LoopVectorize/vplan-sink-scalars-and-merge.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D139788.482414.patch
Type: text/x-patch
Size: 5850 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20221213/618c3452/attachment.bin>