[llvm] [LoopVectorize] Improve Vectorization of Low Trip Count Loops (PR #195823)

[Jack Styles via llvm-commits]

================
@@ -3014,12 +3019,40 @@ LoopVectorizationCostModel::computeMaxVF(ElementCount UserVF, unsigned UserIC) {
       // the trip count but the scalable factor does not, use the fixed-width
       // factor in preference to allow the generation of a non-predicated loop.
       if (EpilogueLoweringStatus == CM_EpilogueNotAllowedLowTripLoop &&
-          NoScalarEpilogueNeeded(MaxFactors.FixedVF.getFixedValue())) {
+          NoScalarEpilogueNeeded(MaxFactors.FixedVF.getFixedValue(),
+                                 EffectiveIC)) {
         LLVM_DEBUG(dbgs() << "LV: Picking a fixed-width so that no tail will "
                              "remain for any chosen VF.\n");
         MaxFactors.ScalableVF = ElementCount::getScalable(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.getFixedValue() != 0) {
+        // If the maximum VF cannot produce 1 vector iteration + 1 scalar
+        // iteration, step down VF's to find one that can. The result should
+        // also eliminate any loops.
+        //
+        // Forced interleaving is considered when seeing if
+        // OneScalarIterationRemainder is produced. It may produced more than
+        // one vector iteration, but only one scalar iteration.
+        for (unsigned MaxVF = MaxFactors.FixedVF.getFixedValue(); MaxVF >= 2;
+             MaxVF /= 2) {
+          if (HasOneScalarIterationRemainder(ExactTC, MaxVF)) {
+            LLVM_DEBUG(dbgs() << "LV: Picking VF=" << MaxVF
+                              << " with 1 scalar iteration remainder.\n");
----------------
Stylie777 wrote:

Done

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