[PATCH] D89650: [DSE] Remove noop stores after killing stores for a MemoryDef.

[Florian Hahn via Phabricator via llvm-commits]

fhahn updated this revision to Diff 301232.
fhahn added a comment.



In D89650#2353901 <https://reviews.llvm.org/D89650#2353901>, @zoecarver wrote:

> This LGTM. I wonder if there are any cases where passing the killing memory def in `storeIsNoop` to `storeIsNoop` would benefit us. I can't think of one off the top of my head, because it doesn't really matter what order the noop stores come in. But maybe if there was a weird block ordering. I'll see if I can come up with an example.

Yes there might be some other cases we still miss with this ordering, but so far from the testing  I did those did not seem to really show up.

We can always further improve this if additional cases surface.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D89650/new/

https://reviews.llvm.org/D89650

Files:
  llvm/lib/Transforms/Scalar/DeadStoreElimination.cpp
  llvm/test/Transforms/DeadStoreElimination/MSSA/calloc-store.ll
  llvm/test/Transforms/DeadStoreElimination/MSSA/noop-stores.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D89650.301232.patch
Type: text/x-patch
Size: 4731 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20201028/50c76677/attachment.bin>