[all-commits] [llvm/llvm-project] fc37f7: [mlir][NVGPU]: Fix op description of nvgpu.device_...

[Yuan Yao via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: fc37f717770acdfe5504bb9b969a01bb16a187f9
      https://github.com/llvm/llvm-project/commit/fc37f717770acdfe5504bb9b969a01bb16a187f9
  Author: Yuan Yao <yuayao at nvidia.com>
  Date:   2023-06-30 (Fri, 30 Jun 2023)

  Changed paths:
    M mlir/include/mlir/Dialect/NVGPU/IR/NVGPU.td

  Log Message:
  -----------
  [mlir][NVGPU]: Fix op description of nvgpu.device_async_wait.

According to the NVIDIA documentation on `cp.async.wait_group`
(https://docs.nvidia.com/cuda/parallel-thread-execution/index.html#data-movement-and-conversion-instructions-cp-async-wait-group-cp-async-wait-all),
the `numGroups` attribute in `nvgpu.device_async_wait` should give an upper
bound of pending async group count (instead of a lower bound) when the
executing thread can be unblocked.

Reviewed By: christopherbate

Differential Revision: https://reviews.llvm.org/D154046