[llvm] [VPlan] Introduce ComputeReductionResult VPInstruction opcode. (PR #70253)

[via llvm-commits]

================
@@ -8970,10 +8848,15 @@ VPlanPtr LoopVectorizationPlanner::buildVPlan(VFRange &Range) {
 // to reductions, with one operand being vector and the other being the scalar
 // reduction chain. For other reductions, a select is introduced between the phi
 // and live-out recipes when folding the tail.
+//
+// A ComputeReductionResult recipe is added to the middle block, also for
+// in-loop reductions which conmpute their result in-loop, because generating
----------------
ayalz wrote:

```suggestion
// in-loop reductions which compute their result in-loop, because generating
```

https://github.com/llvm/llvm-project/pull/70253