[all-commits] [llvm/llvm-project] 1b9b79: [MLIR][Conversion] Convert XeGPU to XeVM pass: Rem...

[Sang Ik Lee via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 1b9b79071de634c8d77785c4d2c8da02857b533b
      https://github.com/llvm/llvm-project/commit/1b9b79071de634c8d77785c4d2c8da02857b533b
  Author: Sang Ik Lee <sang.ik.lee at intel.com>
  Date:   2025-09-10 (Wed, 10 Sep 2025)

  Changed paths:
    M mlir/lib/Conversion/XeGPUToXeVM/XeGPUToXeVM.cpp
    M mlir/test/Conversion/XeGPUToXeVM/create_nd_tdesc.mlir
    M mlir/test/Conversion/XeGPUToXeVM/loadstoreprefetch.mlir
    M mlir/test/Conversion/XeGPUToXeVM/materializecast.mlir
    R mlir/test/Conversion/XeGPUToXeVM/update_offset.mlir

  Log Message:
  -----------
  [MLIR][Conversion] Convert XeGPU to XeVM pass: Remove lowering support for tensor descriptor with offsets. (#157550)

And update load/store/prefetch test cases to use direct offsets.
Tensor descriptors with offsets are getting deprecated.



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