[all-commits] [llvm/llvm-project] f822db: [AMDGPU] Allow for MFMA Inst Clustering

[Jeffrey Byrnes via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: f822db7670d4399bcc90830f23fdb5cec6878c73
      https://github.com/llvm/llvm-project/commit/f822db7670d4399bcc90830f23fdb5cec6878c73
  Author: jeff <jrbyrnes1989 at gmail.com>
  Date:   2022-05-10 (Tue, 10 May 2022)

  Changed paths:
    A llvm/lib/Target/AMDGPU/AMDGPUMFMAClustering.cpp
    A llvm/lib/Target/AMDGPU/AMDGPUMFMAClustering.h
    M llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
    M llvm/lib/Target/AMDGPU/CMakeLists.txt
    A llvm/test/CodeGen/AMDGPU/mfma-cluster-edges.mir
    A llvm/test/CodeGen/AMDGPU/mfma-cluster.mir

  Log Message:
  -----------
  [AMDGPU] Allow for MFMA Inst Clustering

This patch adds cluster edges between independent MFMA instructions. Additionally, it propogates all predecessors of cluster insts to the root of the cluster(s), and all successors to the leaf(ves) of the cluster(s) -- this is done to remove the possibility that those insts will be interspersed within the cluster.

Reviewed By: kerbowa

Differential Revision: https://reviews.llvm.org/D124678