[llvm-branch-commits] [llvm] [SelectionDAG][X86] Split via Concat <n x T> vector types for atomic load (PR #120640)

[via llvm-branch-commits]

================
@@ -5218,7 +5218,11 @@ void SelectionDAGBuilder::visitAtomicLoad(const LoadInst &I) {
     L = DAG.getPtrExtOrTrunc(L, dl, VT);
 
   setValue(&I, L);
-  DAG.setRoot(OutChain);
+
----------------
jofrn wrote:

It is here so that the elements are vectorized later in EltsFromConsecutiveLoads. Without this here, it will need to check the chain differently in EltsFromConsecutiveLoads based on being an Atomic or not in order to vectorize.

https://github.com/llvm/llvm-project/pull/120640