[all-commits] [llvm/llvm-project] a6a921: Lower shuffle to single-result form if possible. (...

[Johannes Reifferscheid via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: a6a9215b93bcbf901cd11d2dd02cce1a245d3ffe
      https://github.com/llvm/llvm-project/commit/a6a9215b93bcbf901cd11d2dd02cce1a245d3ffe
  Author: Johannes Reifferscheid <jreiffers at google.com>
  Date:   2024-03-21 (Thu, 21 Mar 2024)

  Changed paths:
    M mlir/lib/Conversion/GPUToNVVM/LowerGpuOpsToNVVMOps.cpp
    M mlir/test/Conversion/GPUToNVVM/gpu-to-nvvm.mlir

  Log Message:
  -----------
  Lower shuffle to single-result form if possible. (#84321)

We currently always lower shuffle to the struct-returning variant. I saw
some cases where this survived all the way through ptx, resulting in
increased register usage. The easiest fix is to simply lower to the
single-result version when the predicate is unused.



To unsubscribe from these emails, change your notification settings at https://github.com/llvm/llvm-project/settings/notifications