[all-commits] [llvm/llvm-project] fb2c42: [OpenMP] Improve AMDGPU Plugin

[Johannes Doerfert via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: fb2c42df41cb01e1122fd4e9c81e1f4bc5592b12
      https://github.com/llvm/llvm-project/commit/fb2c42df41cb01e1122fd4e9c81e1f4bc5592b12
  Author: Johannes Doerfert <johannes at jdoerfert.de>
  Date:   2022-12-19 (Mon, 19 Dec 2022)

  Changed paths:
    M llvm/include/llvm/Frontend/OpenMP/OMPGridValues.h
    M openmp/libomptarget/plugins-nextgen/amdgpu/src/rtl.cpp
    M openmp/libomptarget/plugins-nextgen/common/PluginInterface/PluginInterface.cpp
    M openmp/libomptarget/plugins-nextgen/common/PluginInterface/PluginInterface.h

  Log Message:
  -----------
  [OpenMP] Improve AMDGPU Plugin

With this patch we:
- pick more sensible defaults for the number of teams, inspired by the
  old plugin, and configured via LIBOMPTARGET_AMDGPU_TEAMS_PER_CU.
- check the input signal of a kernel launch late, after the queue lock
  was taken, to avoid a barrier packet more often.
- copy the kernel arguments in one swoop into the appropriate memory.
- manually specialize the callbacks to avoid potential indirect calls.


  Commit: e3d9a448c57c46600db635244792205f7a0ea58b
      https://github.com/llvm/llvm-project/commit/e3d9a448c57c46600db635244792205f7a0ea58b
  Author: Johannes Doerfert <johannes at jdoerfert.de>
  Date:   2022-12-19 (Mon, 19 Dec 2022)

  Changed paths:
    M openmp/libomptarget/plugins-nextgen/amdgpu/src/rtl.cpp

  Log Message:
  -----------
  [OpenMP] Account for dynamic shared memory in the AMDGPU nextgen plugin


Compare: https://github.com/llvm/llvm-project/compare/ab17a08d0f6f...e3d9a448c57c