[VPlan] Replace AlsoPack field with shouldPack() method (NFC).

[fhahn]

Summary

There is no need to update the AlsoPack field when creating
VPReplicateRecipes. It can be easily computed based on the VP def-use
chains when it is needed.

[fhahn]

ping :)

[Ayal]

Indeed better compute such information on demand than cache it! Looks good to me, adding a couple of nits.

llvm/lib/Transforms/Vectorize/LoopVectorize.cpp

9573

nit: may be slightly faster if swapped to check State.VF.isVector() first.

llvm/lib/Transforms/Vectorize/VPlan.h

1546

nit: used by a widened instruction indirectly - via an intervening PredInstPhi? Until the insert-element's are represented explicitly.

llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp

929

This is indeed the current behavior, but better have it check if !U.usesScalars() instead, possibly as a follow-up?

[fhahn]

Rebase and address comments, thanks! I plan to land this shortly.

llvm/lib/Transforms/Vectorize/LoopVectorize.cpp

if (!IsUniform && Range.Start.isScalable() && isa<IntrinsicInst>(I)) {
     default:
       break;
     }
   }
 
   auto *Recipe = new VPReplicateRecipe(I, Plan.mapToVPValues(I->operands()),
                                        IsUniform, IsPredicated);
 
-  // Find if I uses a predicated instruction. If so, it will use its scalar
-  // value. Avoid hoisting the insert-element which packs the scalar value into
-  // a vector value, as that happens iff all users use the vector value.
-  for (VPValue *Op : Recipe->operands()) {
-    auto *PredR =
-        dyn_cast_or_null<VPPredInstPHIRecipe>(Op->getDefiningRecipe());
-    if (!PredR)
-      continue;
-    auto *RepR = cast<VPReplicateRecipe>(
-        PredR->getOperand(0)->getDefiningRecipe());
-    assert(RepR->isPredicated() &&
-           "expected Replicate recipe to be predicated");
-    RepR->setAlsoPack(false);
-  }
-
   // Finalize the recipe for Instr, first if it is not predicated.
   if (!IsPredicated) {
     LLVM_DEBUG(dbgs() << "LV: Scalarizing:" << *I << "\n");
     setRecipe(I, Recipe);
     Plan.addVPValue(I, Recipe);
     VPBB->appendRecipe(Recipe);
     return VPBB;
   }

 
 void VPReplicateRecipe::execute(VPTransformState &State) {
   Instruction *UI = getUnderlyingInstr();
   if (State.Instance) { // Generate a single instance.
     assert(!State.VF.isScalable() && "Can't scalarize a scalable vector");
     State.ILV->scalarizeInstruction(UI, this, *State.Instance,
                                     IsPredicated, State);
     // Insert scalar instance packing it into a vector.
-    if (AlsoPack && State.VF.isVector()) {
+    if (State.VF.isVector() && shouldPack()) {
       // If we're constructing lane 0, initialize to start from poison.
       if (State.Instance->Lane.isFirstLane()) {
         assert(!State.VF.isScalable() && "VF is assumed to be non scalable.");
         Value *Poison = PoisonValue::get(
             VectorType::get(UI->getType(), State.VF));
         State.set(this, Poison, State.Instance->Part);
       }
       State.ILV->packScalarIntoVectorValue(this, *State.Instance, State);

llvm/lib/Transforms/Vectorize/VPRecipeBuilder.h

public:
 
   /// Create a replicating region for \p PredRecipe.
   VPRegionBlock *createReplicateRegion(VPReplicateRecipe *PredRecipe,
                                        VPlan &Plan);
 
   /// Build a VPReplicationRecipe for \p I and enclose it within a Region if it
   /// is predicated. \return \p VPBB augmented with this new recipe if \p I is
   /// not predicated, otherwise \return a new VPBasicBlock that succeeds the new
-  /// Region. Update the packing decision of predicated instructions if they
-  /// feed \p I. Range.End may be decreased to ensure same recipe behavior from
-  /// \p Range.Start to \p Range.End.
+  /// Region. Range.End may be decreased to ensure same recipe behavior from \p
+  /// Range.Start to \p Range.End.
   VPBasicBlock *handleReplication(Instruction *I, VFRange &Range,
                                   VPBasicBlock *VPBB, VPlan &Plan);
 
   /// Add the incoming values from the backedge to reduction & first-order
   /// recurrence cross-iteration phis.
   void fixHeaderPhis();
 };
 } // end namespace llvm
 
 #endif // LLVM_TRANSFORMS_VECTORIZE_VPRECIPEBUILDER_H

llvm/lib/Transforms/Vectorize/VPlan.h

 /// uniform only one copy, per lane zero, will be generated.
 class VPReplicateRecipe : public VPRecipeBase, public VPValue {
   /// Indicator if only a single replica per lane is needed.
   bool IsUniform;
 
   /// Indicator if the replicas are also predicated.
   bool IsPredicated;
 
-  /// Indicator if the scalar values should also be packed into a vector.
-  bool AlsoPack;
-
 public:
   template <typename IterT>
   VPReplicateRecipe(Instruction *I, iterator_range<IterT> Operands,
                     bool IsUniform, bool IsPredicated = false)
       : VPRecipeBase(VPDef::VPReplicateSC, Operands), VPValue(this, I),
-        IsUniform(IsUniform), IsPredicated(IsPredicated) {
-    // Retain the previous behavior of predicateInstructions(), where an
-    // insert-element of a predicated instruction got hoisted into the
-    // predicated basic block iff it was its only user. This is achieved by
-    // having predicated instructions also pack their values into a vector by
-    // default unless they have a replicated user which uses their scalar value.
-    AlsoPack = IsPredicated && !I->use_empty();
-  }
+        IsUniform(IsUniform), IsPredicated(IsPredicated) {}
 
   ~VPReplicateRecipe() override = default;
 
   VP_CLASSOF_IMPL(VPDef::VPReplicateSC)
 
   /// Generate replicas of the desired Ingredient. Replicas will be generated
   /// for all parts and lanes unless a specific part and lane are specified in
   /// the \p State.
   void execute(VPTransformState &State) override;
 
-  void setAlsoPack(bool Pack) { AlsoPack = Pack; }
-
 #if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
   /// Print the recipe.
   void print(raw_ostream &O, const Twine &Indent,
              VPSlotTracker &SlotTracker) const override;
 #endif
 
   bool isUniform() const { return IsUniform; }
 
-  bool isPacked() const { return AlsoPack; }
-
   bool isPredicated() const { return IsPredicated; }
 
   /// Returns true if the recipe only uses the first lane of operand \p Op.
   bool onlyFirstLaneUsed(const VPValue *Op) const override {
     assert(is_contained(operands(), Op) &&
            "Op must be an operand of the recipe");
     return isUniform();
   }
 
   /// Returns true if the recipe uses scalars of operand \p Op.
   bool usesScalars(const VPValue *Op) const override {
     assert(is_contained(operands(), Op) &&
            "Op must be an operand of the recipe");
     return true;
   }
+
+  /// Returns true if the recipe is used by a widened recipe via an intervening
+  /// VPPredInstPHIRecipe. In this case, the scalar values should also be packed
+  /// in a vector.
+  bool shouldPack() const;
 };
 
 /// A recipe for generating conditional branches on the bits of a mask.
 class VPBranchOnMaskRecipe : public VPRecipeBase {
 public:
   VPBranchOnMaskRecipe(VPValue *BlockInMask)
       : VPRecipeBase(VPDef::VPBranchOnMaskSC, {}) {
     if (BlockInMask) // nullptr means all-one mask.

llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp

if (getCondOp()) {
     getCondOp()->printAsOperand(O, SlotTracker);
   }
   O << ")";
   if (RdxDesc->IntermediateStore)
     O << " (with final reduction value stored in invariant address sank "
          "outside of loop)";
 }
 
+bool VPReplicateRecipe::shouldPack() const {
+  // Find if the recipe is used by a widened recipe via an intervening
+  // VPPredInstPHIRecipe. In this case, also pack the scalar values in a vector.
+  return any_of(users(), [](const VPUser *U) {
+    if (auto *PredR = dyn_cast<VPPredInstPHIRecipe>(U))
+      return any_of(PredR->users(),
+                    [](const VPUser *U) { return !isa<VPReplicateRecipe>(U); });
+    return false;
+  });
+}
 void VPReplicateRecipe::print(raw_ostream &O, const Twine &Indent,
                               VPSlotTracker &SlotTracker) const {
   O << Indent << (IsUniform ? "CLONE " : "REPLICATE ");
 
   if (!getUnderlyingInstr()->getType()->isVoidTy()) {
     printAsOperand(O, SlotTracker);
     O << " = ";
   }
   if (auto *CB = dyn_cast<CallBase>(getUnderlyingInstr())) {
     O << "call @" << CB->getCalledFunction()->getName() << "(";
     interleaveComma(make_range(op_begin(), op_begin() + (getNumOperands() - 1)),
                     O, [&O, &SlotTracker](VPValue *Op) {
                       Op->printAsOperand(O, SlotTracker);
                     });
     O << ")";
   } else {
     O << Instruction::getOpcodeName(getUnderlyingInstr()->getOpcode()) << " ";
     printOperands(O, SlotTracker);
   }
 
-  if (AlsoPack)
+  if (shouldPack())
     O << " (S->V)";
 }
 #endif
 
 void VPBranchOnMaskRecipe::execute(VPTransformState &State) {
   assert(State.Instance && "Branch on Mask works only on single instance.");
 
   unsigned Part = State.Instance->Part;