[all-commits] [llvm/llvm-project] 38049d: [libc] Handle differing wavefront sizes correctly ...

[Joseph Huber via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 38049dc8eef0dca7e82c25e4012228a9a135e255
      https://github.com/llvm/llvm-project/commit/38049dc8eef0dca7e82c25e4012228a9a135e255
  Author: Joseph Huber <huberjn at outlook.com>
  Date:   2024-11-27 (Wed, 27 Nov 2024)

  Changed paths:
    M libc/cmake/modules/LLVMLibCTestRules.cmake
    M libc/test/integration/startup/gpu/CMakeLists.txt
    A libc/test/integration/startup/gpu/rpc_lane_test.cpp
    M libc/utils/gpu/loader/amdgpu/CMakeLists.txt
    M libc/utils/gpu/loader/amdgpu/amdhsa-loader.cpp

  Log Message:
  -----------
  [libc] Handle differing wavefront sizes correctly in the AMDHSA loader (#117788)

Summary:
The AMDGPU backend can handle wavefront sizes of 32 and 64, with the
native hardware preferring one or the other. The user can override the
hardware with `-mwavefrontsize64` or `-mwavefrontsize32` which
previously wasn't handled. We need to know the wavefront size to know
how much memory to allocate and how to index the RPC buffer. There isn't
a good way to do this with ROCm so we just use the LLVM support for
offloading to check this from the image.



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