[llvm] f8bbb21 - AMDGPU: Handle invariant when lowering global loads (#168914)

[via llvm-commits]

Author: Matt Arsenault
Date: 2025-11-21T13:58:35-05:00
New Revision: f8bbb21fb9d25c18cad6e197eb29378f060975aa

URL: https://github.com/llvm/llvm-project/commit/f8bbb21fb9d25c18cad6e197eb29378f060975aa
DIFF: https://github.com/llvm/llvm-project/commit/f8bbb21fb9d25c18cad6e197eb29378f060975aa.diff

LOG: AMDGPU: Handle invariant when lowering global loads (#168914)

Global with invariant should be treated identically to
constant.

Added: 
    

Modified: 
    llvm/lib/Target/AMDGPU/SIISelLowering.cpp

Removed: 
    


################################################################################
diff  --git a/llvm/lib/Target/AMDGPU/SIISelLowering.cpp b/llvm/lib/Target/AMDGPU/SIISelLowering.cpp
index 875278a3b4f97..c681d12ba7499 100644
--- a/llvm/lib/Target/AMDGPU/SIISelLowering.cpp
+++ b/llvm/lib/Target/AMDGPU/SIISelLowering.cpp
@@ -11944,7 +11944,7 @@ SDValue SITargetLowering::LowerLOAD(SDValue Op, SelectionDAG &DAG) const {
       AS == AMDGPUAS::CONSTANT_ADDRESS_32BIT ||
       (AS == AMDGPUAS::GLOBAL_ADDRESS &&
        Subtarget->getScalarizeGlobalBehavior() && Load->isSimple() &&
-       isMemOpHasNoClobberedMemOperand(Load))) {
+       (Load->isInvariant() || isMemOpHasNoClobberedMemOperand(Load)))) {
     if ((!Op->isDivergent() || AMDGPU::isUniformMMO(MMO)) &&
         Alignment >= Align(4) && NumElements < 32) {
       if (MemVT.isPow2VectorType() ||