[all-commits] [llvm/llvm-project] 9f69d3: [Libomptarget] Use NVPTX lane id intrinsic in Devi...

[Joseph Huber via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 9f69d3cf88905df5006f93dce536b7e73c0b1735
      https://github.com/llvm/llvm-project/commit/9f69d3cf88905df5006f93dce536b7e73c0b1735
  Author: Joseph Huber <huberjn at outlook.com>
  Date:   2024-03-12 (Tue, 12 Mar 2024)

  Changed paths:
    M openmp/libomptarget/DeviceRTL/src/Mapping.cpp

  Log Message:
  -----------
  [Libomptarget] Use NVPTX lane id intrinsic in DeviceRTL (#84928)

Summary:
We are currently taking the lower 5 bites of the thread ID as the warp
ID. This doesn't work in non-1D grids and is also slower than just using
the dedicated hardware register.



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