[llvm] [LoopVectorize] Improve Vectorization of Small Loops (PR #195823)

[Jack Styles via llvm-commits]

================
@@ -3023,9 +3033,37 @@ LoopVectorizationCostModel::computeMaxVF(ElementCount UserVF, unsigned UserIC) {
           NoScalarEpilogueNeeded(MaxFactors.FixedVF.getFixedValue())) {
         LLVM_DEBUG(dbgs() << "LV: Picking a fixed-width so that no tail will "
                              "remain for any chosen VF.\n");
-        MaxFactors.ScalableVF = ElementCount::getScalable(0);
+        ApplyVectorWidth(MaxFactors, MaxFactors.FixedVF.getFixedValue(), 0);
         return MaxFactors;
       }
+      // Allow cases where the ExactTC == VF + 1. VF can be any power of
+      // 2 between 2 and MaxVF.
+      //
+      // This produces 1 vector iteration, and 1 scalar iteration with
+      // no remainder. Later passes will eliminate the loop and leave
+      // straight-line code as the both iteration counts are statically known.
+      ElementCount ExactTC = getSmallConstantTripCount(PSE.getSE(), TheLoop);
+      if (EpilogueLoweringStatus == CM_EpilogueNotAllowedLowTripLoop &&
+          ExactTC && ExactTC.isFixed()) {
----------------
Stylie777 wrote:

If ExpectedTC is fixed, then ExactTC will also be fixed. I've changed this to just check ExactTC.isFixed()

https://github.com/llvm/llvm-project/pull/195823