[llvm] [LV][EVL] Support fixed-order recurrence idiom with EVL tail folding. (PR #124093)

[Luke Lau via llvm-commits]

================
@@ -1727,6 +1783,30 @@ static void transformRecipestoEVLRecipes(VPlan &Plan, VPValue &EVL) {
   VPTypeAnalysis TypeInfo(CanonicalIVType);
   LLVMContext &Ctx = CanonicalIVType->getContext();
   VPValue *AllOneMask = Plan.getOrAddLiveIn(ConstantInt::getTrue(Ctx));
+  VPRegionBlock *LoopRegion = Plan.getVectorLoopRegion();
+  VPBasicBlock *Header = LoopRegion->getEntryBasicBlock();
+
+  // Create a scalar phi to track the previous EVL if fixed-order recurrence is
+  // contained.
+  VPScalarPHIRecipe *PrevEVL = nullptr;
+  bool ContainsFORs =
+      any_of(Header->phis(), IsaPred<VPFirstOrderRecurrencePHIRecipe>);
+  if (ContainsFORs) {
+    // TODO: Use VPInstruction::ExplicitVectorLength to get maximum EVL.
----------------
lukel97 wrote:

Oh ok I see how that would work if we change `VPFirstOrderRecurrencePHIRecipe` too to not use getRuntimeVF. 

I think what the current PR has is good so I'll mark this as resolved. I'm not sure about what you mean about the performance though. Would using ExplicitVectorLength instead of VF help performance?

https://github.com/llvm/llvm-project/pull/124093