[all-commits] [llvm/llvm-project] 85ec95: [AMDGPU] A Vulkan-style memory model weaker than t...

[Sameer Sahasrabuddhe via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 85ec9552dc84cbc5acd025bdeb9b258b044591b9
      https://github.com/llvm/llvm-project/commit/85ec9552dc84cbc5acd025bdeb9b258b044591b9
  Author: Sameer Sahasrabuddhe <sameer.sahasrabuddhe at amd.com>
  Date:   2026-06-01 (Mon, 01 Jun 2026)

  Changed paths:
    A llvm/docs/AMDGPUMemoryModel.rst
    M llvm/docs/AMDGPUUsage.rst
    M llvm/docs/UserGuides.rst

  Log Message:
  -----------
  [AMDGPU] A Vulkan-style memory model weaker than the LLVM model (#191246)

Add a new AMDGPU memory model specification that is weaker than the LLVM
memory model using Vulkan-style availability/visibility semantics and
scoped operations. The model allows more efficient implementations while
maintaining a safe-by-default mapping to the standard LLVM model.

**RFC:**
https://discourse.llvm.org/t/a-vulkan-style-memory-model-for-amdgpu-and-beyond/90498

Related changes:

- #191390
- #199489



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