[llvm] [LV] Add cross-part load overlap heuristic for interleaving (PR #214500)

[Sergey Shcherbinin via llvm-commits]

================
@@ -3630,9 +3649,53 @@ std::unique_ptr<VPlan> LoopVectorizationPlanner::selectBestEpiloguePlan(
   return Clone;
 }
 
+bool LoopVectorizationPlanner::shouldUseCrossPartCSE() const {
+  // Keep the disabled path free of policy queries and cost-map allocation.
+  if (!EnableInterleaveCSE)
+    return false;
+  return OrigLoop->isInnermost() && Config.getHints().getInterleave() == 0 &&
+         !CM->maskPartialAliasing();
+}
+
+bool LoopVectorizationPlanner::shouldCollectCrossPartCSECosts(
+    VPlan &Plan, ElementCount VF) const {
+  return shouldUseCrossPartCSE() && VF.isVector() && !VF.isScalable() &&
----------------
SergeyShch01 wrote:

 The restriction is now removed, with positive and negative scalable forward/reverse tests added.

https://github.com/llvm/llvm-project/pull/214500