[llvm] [NVPTX] Optimize v16i8 reductions (PR #67322)

[Artem Belevich via llvm-commits]

================
@@ -5294,6 +5295,98 @@ static SDValue PerformEXTRACTCombine(SDNode *N,
   return Result;
 }
 
+static SDValue PerformLOADCombine(SDNode *N,
+                                  TargetLowering::DAGCombinerInfo &DCI) {
+  SelectionDAG &DAG = DCI.DAG;
+  LoadSDNode *LD = cast<LoadSDNode>(N);
+
+  // Lower a v16i8 load into a LoadV4 operation with i32 results instead of
+  // letting ReplaceLoadVector split it into smaller loads during legalization.
+  // This is done at dag-combine1 time, so that vector operations with i8
+  // elements can be optimised away instead of being needlessly split during
+  // legalization, which involves storing to the stack and loading it back.
+  EVT VT = N->getValueType(0);
+  if (VT != MVT::v16i8)
----------------
Artem-B wrote:

Do we want to generalize it to v8i8, too?

https://github.com/llvm/llvm-project/pull/67322