[Mlir-commits] [mlir] Added Memory Effect to assume_alignment (PR #139521)

[llvmlistbot at llvm.org]

https://github.com/MaheshRavishankar requested changes to this pull request.

Saying it has write semantics is a big pessimization. This is more  a compiler hint. If an optimization is removing it, that needs to be fixed in the optimization AFAICS. For example CSE does no CSE operation with no results (even if it has no side-effects).

https://github.com/llvm/llvm-project/pull/139521