aemerson wrote: I managed to reduce down the original SME test to `Transforms/PhaseOrdering/always-inline-alloca-promotion.ll`. Compiling that to assembly with clang with and without the change shows the differences in codegen quality, but the IR shows the kind of scenario this patch is meant to handle.
https://github.com/llvm/llvm-project/pull/145613 _______________________________________________ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits