[llvm] [InstCombine] Fold binary op of reductions. (PR #121567)

[Mikhail Gudim via llvm-commits]

================
@@ -2313,6 +2313,62 @@ Instruction *InstCombinerImpl::foldVectorBinop(BinaryOperator &Inst) {
   return nullptr;
 }
 
+static Intrinsic::ID getReductionForBinop(Instruction::BinaryOps Opc) {
+  switch (Opc) {
+  default:
+    break;
+  case Instruction::Add:
----------------
mgudim wrote:

> You could also fold or disjoint with vector_reduce_add

what would be the benefit of doing this? 

https://github.com/llvm/llvm-project/pull/121567