[all-commits] [llvm/llvm-project] 5f6634: Use nvptxcompile library.

[Artem Belevich via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 5f66348e59aa7ce5e5780a972b3875268c45d57c
      https://github.com/llvm/llvm-project/commit/5f66348e59aa7ce5e5780a972b3875268c45d57c
  Author: Artem Belevich <tra at google.com>
  Date:   2023-03-17 (Fri, 17 Mar 2023)

  Changed paths:
    M mlir/lib/Dialect/GPU/CMakeLists.txt
    M mlir/lib/Dialect/GPU/Transforms/SerializeToCubin.cpp
    M mlir/lib/ExecutionEngine/CMakeLists.txt

  Log Message:
  -----------
  Use nvptxcompile library.

Differential Revision: https://reviews.llvm.org/D145527


  Commit: fea8faa705b43e23508bc9fef842c62929786110
      https://github.com/llvm/llvm-project/commit/fea8faa705b43e23508bc9fef842c62929786110
  Author: Artem Belevich <tra at google.com>
  Date:   2023-03-17 (Fri, 17 Mar 2023)

  Changed paths:
    M utils/bazel/llvm-project-overlay/mlir/BUILD.bazel

  Log Message:
  -----------
  Separate SerializeToCubin from GPUTransforms.

SerializeToCubin depends on CUDA at *runtime* which is undesirable for MLIR's
general use case, as compilation should be doable on any host, regardless of
whether it has a GPU.

SerializeToCubin is needed to run some GPU tests, so when we build mlir-opt,
SerializeToCubin pass is linked in directly into mlir-opt.

Differential Revision: https://reviews.llvm.org/D146330


Compare: https://github.com/llvm/llvm-project/compare/1c9ec74e3f2a...fea8faa705b4