[llvm] [AMDGPU] Cluster export instruction in PostRA Scheduler (PR #141399)

[via llvm-commits]

llvmbot wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-backend-amdgpu

Author: Carl Ritson (perlfu)

<details>
<summary>Changes</summary>

DAG mutation needs to be applied post-RA to maintain order established during pre-RA scheduler.

---
Full diff: https://github.com/llvm/llvm-project/pull/141399.diff


1 Files Affected:

- (modified) llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp (+1) 


``````````diff
diff --git a/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp b/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
index e24d8481408ad..7479703ce353a 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
@@ -1129,6 +1129,7 @@ GCNTargetMachine::createPostMachineScheduler(MachineSchedContext *C) const {
        getOptLevel() >= CodeGenOptLevel::Less) &&
       EnableVOPD)
     DAG->addMutation(createVOPDPairingMutation());
+  DAG->addMutation(createAMDGPUExportClusteringDAGMutation());
   return DAG;
 }
 //===----------------------------------------------------------------------===//

``````````

</details>


https://github.com/llvm/llvm-project/pull/141399