[llvm] [VPlan] Move call widening decision to VPlan. (NFCI) (PR #195518)

[Ramkumar Ramachandra via llvm-commits]

================
@@ -1855,6 +1856,11 @@ void VPWidenCallRecipe::execute(VPTransformState &State) {
 
 InstructionCost VPWidenCallRecipe::computeCost(ElementCount VF,
                                                VPCostContext &Ctx) const {
+  return computeVectorCallCost(Variant, Ctx);
+}
+
+InstructionCost VPWidenCallRecipe::computeVectorCallCost(Function *Variant,
+                                                         VPCostContext &Ctx) {
----------------
artagnon wrote:

Hm, interesting. It doesn't depend on VF, but may be good to add an assert that the Variant's return type is VF wide?

https://github.com/llvm/llvm-project/pull/195518