[all-commits] [llvm/llvm-project] 065179: [mlir][OpenACC] Lower single-block thread-only arr...

Matsu via All-commits all-commits at lists.llvm.org
Mon Jul 27 21:21:24 PDT 2026


  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 065179637e5331424ffdef8060593c57a53a6391
      https://github.com/llvm/llvm-project/commit/065179637e5331424ffdef8060593c57a53a6391
  Author: Matsu <47756807+khaki3 at users.noreply.github.com>
  Date:   2026-07-27 (Mon, 27 Jul 2026)

  Changed paths:
    M mlir/lib/Dialect/OpenACC/Transforms/ACCCGToGPU.cpp
    M mlir/test/Dialect/OpenACC/acc-cg-to-gpu-reduction-array.mlir

  Log Message:
  -----------
  [mlir][OpenACC] Lower single-block thread-only array reductions (#212369)

Example:
```fortran
!$acc parallel loop vector reduction(+:b)
do i = 1, n
  b(i) = i
end do
```

In this code the array reduction is thread(vector)-only with no
gang/block launch dim, so a within-block `all_reduce` is a complete
reduction. ACCCGToGPU still rejected it as NYI because the accumulate
had no block context.

Fix: keep the NYI only when the region actually launches block
dimensions; allow the existing per-element `all_reduce` path for
single-block thread-only array accumulates.



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


More information about the All-commits mailing list