[all-commits] [llvm/llvm-project] 652ff2: [flang][cuda] Adding atomicadd as a cudadevice int...

[Renaud Kauffmann via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 652ff20140d79544db4dfa21314fc62c3c9182e5
      https://github.com/llvm/llvm-project/commit/652ff20140d79544db4dfa21314fc62c3c9182e5
  Author: Renaud Kauffmann <rkauffmann at nvidia.com>
  Date:   2025-01-22 (Wed, 22 Jan 2025)

  Changed paths:
    M flang/include/flang/Optimizer/Builder/IntrinsicCall.h
    M flang/lib/Optimizer/Builder/IntrinsicCall.cpp
    M flang/module/cudadevice.f90
    M flang/test/Lower/CUDA/cuda-device-proc.cuf
    M flang/test/Semantics/cuf-device-procedures01.cuf

  Log Message:
  -----------
  [flang][cuda] Adding atomicadd as a cudadevice intrinsic and converting it LLVM dialect (#123840)

With these changes, CUF atomic operations are handled as cudadevice
intrinsics and are converted straight to the LLVM dialect with the
`llvm.atomicrw` operation.

I am only submitting changes for `atomicadd` to gather feedback. If we
are to proceed with these changes I will add support for all other
applicable atomic operations following this pattern.



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