[llvm] Scalarize the vector inputs to llvm.lround intrinsic by default. (PR #101054)

[Matt Arsenault via llvm-commits]

================
@@ -17856,15 +17858,17 @@ SDValue DAGCombiner::visitFP_TO_UINT(SDNode *N) {
   return FoldIntToFPToInt(N, DAG);
 }
 
-SDValue DAGCombiner::visitXRINT(SDNode *N) {
+SDValue DAGCombiner::visitXROUND(SDNode *N) {
   SDValue N0 = N->getOperand(0);
   EVT VT = N->getValueType(0);
 
   // fold (lrint|llrint undef) -> undef
+  // fold (lround|llround undef) -> undef
----------------
arsenm wrote:

I don't think this is tested. This *might* happen for the v3f16 case 

https://github.com/llvm/llvm-project/pull/101054