[all-commits] [llvm/llvm-project] e892f3: [mlir][nvgpu] Allow TMA's last dim to be non-128B ...

[Guray Ozen via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: e892f323ea234a997888b0fa44356b0678eda5f0
      https://github.com/llvm/llvm-project/commit/e892f323ea234a997888b0fa44356b0678eda5f0
  Author: Guray Ozen <guray.ozen at gmail.com>
  Date:   2024-02-13 (Tue, 13 Feb 2024)

  Changed paths:
    M mlir/lib/Dialect/NVGPU/IR/NVGPUDialect.cpp

  Log Message:
  -----------
  [mlir][nvgpu] Allow TMA's last dim to be non-128B without swizzling (#81499)

Allow TMA's last dimension to be non-128B when swizzling mode is not
set.

Test `tma_load_64x8_8x128_noswizzle.mlir` is failing due to the
verifier. This PR will fix that