[all-commits] [llvm/llvm-project] 8c9e0c: [flang][OpenMP] Allocate `reduction` init temps on...

[Kareem Ergawy via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 8c9e0c6c61f653928a992422d534e4e7f976dd55
      https://github.com/llvm/llvm-project/commit/8c9e0c6c61f653928a992422d534e4e7f976dd55
  Author: Kareem Ergawy <kareem.ergawy at amd.com>
  Date:   2025-07-04 (Fri, 04 Jul 2025)

  Changed paths:
    M flang/lib/Lower/Support/PrivateReductionUtils.cpp
    M flang/test/Lower/OpenMP/parallel-reduction-array.f90
    M mlir/lib/Target/LLVMIR/Dialect/OpenMP/OpenMPToLLVMIRTranslation.cpp

  Log Message:
  -----------
  [flang][OpenMP] Allocate `reduction` init temps on the stack for GPUs (#146667)

Temps needed for the reduction init regions are now allocate on the heap
all the time. However, this is performance killer for GPUs since malloc
calls are prohibitively expensive. Therefore, we should do these
allocations on the stack for GPU reductions.



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