[LV][VPlan] Add initial support for CSA vectorization

This patch adds initial support for CSA vectorization LLVM. This new class
can be characterized by vectorization of assignment to a scalar in a loop,
such that the assignment is conditional from the perspective of its use.
An assignment is conditional in a loop if a value may or may not be assigned
in the loop body.

For example:

```
int t = init_val;
for (int i = 0; i < N; i++) {
  if (cond[i])
    t = a[i];
}
s = t; // use t
```

Using pseudo-LLVM code this can be vectorized as

```
vector.ph:
  ...
  %t = %init_val
  %init.mask = <all-false-vec>
  %init.data = <poison-vec> ; uninitialized
vector.body:
  ...
  %mask.phi = phi [%init.mask, %vector.ph], [%new.mask, %vector.body]
  %data.phi = phi [%data.mask, %vector.ph], [%new.mask, %vector.body]
  %cond.vec = <widened-cmp> ...
  %a.vec    = <widened-load> %a, %i
  %b        = <any-lane-active> %cond.vec
  %new.mask = select %b, %cond.vec, %mask.phi
  %new.data = select %b, %a.vec, %data.phi
  ...
middle.block:
  %s = <extract-last-active-lane> %new.mask, %new.data
```

On each iteration, we track whether any lane in the widened condition was active,
and if it was take the current mask and data as the new mask and data vector.
Then at the end of the loop, the scalar can be extracted only once.

This transformation works the same way for integer, pointer, and floating point
conditional assignment, since the transformation does not require inspection
of the data being assigned.

In the vectorization of a CSA, we will be introducing recipes into the vector
preheader, the vector body, and the middle block. Recipes that are introduced
into the preheader and middle block are executed only one time, and recipes
that are in the vector body will be possibly executed multiple times. The more
times that the vector body is executed, the less of an impact the preheader
and middle block cost have on the overall cost of a CSA.

A detailed explanation of the concept can be found [here](https://discourse.llvm.org/t/vectorization-of-conditional-scalar-assignment-csa/80964).

This patch is further tested in llvm/llvm-test-suite#155.

This patch contains only the non-EVL related code. The is based on the larger
patch of llvm#106560, which contained both EVL and non-EVL related parts.

Author: michaelmaitland

llvm/include/llvm/Analysis/IVDescriptors.h

@@ -6,7 +6,8 @@
 //
 //===----------------------------------------------------------------------===//
 //
-// This file "describes" induction and recurrence variables.
+// This file "describes" induction, recurrence, and conditional scalar
+// assignment (CSA) variables.
 //
 //===----------------------------------------------------------------------===//
 
@@ -423,6 +424,61 @@ class InductionDescriptor {
   SmallVector<Instruction *, 2> RedundantCasts;
 };
 
+/// A Conditional Scalar Assignment (CSA) is an assignment from an initial
+/// scalar that may or may not occur.
+class CSADescriptor {
+  /// If the conditional assignment occurs inside a loop, then Phi chooses
+  /// the value of the assignment from the entry block or the loop body block.
+  PHINode *Phi = nullptr;
+
+  /// The initial value of the CSA. If the condition guarding the assignment is
+  /// not met, then the assignment retains this value.
+  Value *InitScalar = nullptr;
+
+  /// The Instruction that conditionally assigned to inside the loop.
+  Instruction *Assignment = nullptr;
+
+  /// Create a CSA Descriptor that models a valid CSA with its members
+  /// initialized correctly.
+  CSADescriptor(PHINode *Phi, Instruction *Assignment, Value *InitScalar)
+      : Phi(Phi), InitScalar(InitScalar), Assignment(Assignment) {}
+
+public:
+  /// Create a CSA Descriptor that models an invalid CSA.
+  CSADescriptor() = default;
+
+  /// If Phi is the root of a CSA, set CSADesc as the CSA rooted by
+  /// Phi. Otherwise, return a false, leaving CSADesc unmodified.
+  static bool isCSAPhi(PHINode *Phi, Loop *TheLoop, CSADescriptor &CSADesc);
+
+  operator bool() const { return isValid(); }
+
+  /// Returns whether SI is the Assignment in CSA
+  static bool isCSASelect(CSADescriptor Desc, SelectInst *SI) {
+    return Desc.getAssignment() == SI;
+  }
+
+  /// Return whether this CSADescriptor models a valid CSA.
+  bool isValid() const { return Phi && InitScalar && Assignment; }
+
+  /// Return the PHI that roots this CSA.
+  PHINode *getPhi() const { return Phi; }
+
+  /// Return the initial value of the CSA. This is the value if the conditional
+  /// assignment does not occur.
+  Value *getInitScalar() const { return InitScalar; }
+
+  /// The Instruction that is used after the loop
+  Instruction *getAssignment() const { return Assignment; }
+
+  /// Return the condition that this CSA is conditional upon.
+  Value *getCond() const {
+    if (auto *SI = dyn_cast_or_null<SelectInst>(Assignment))
+      return SI->getCondition();
+    return nullptr;
+  }
+};
+
 } // end namespace llvm
 
 #endif // LLVM_ANALYSIS_IVDESCRIPTORS_H

llvm/include/llvm/Analysis/TargetTransformInfo.h

@@ -1834,6 +1834,10 @@ class TargetTransformInfo {
         : EVLParamStrategy(EVLParamStrategy), OpStrategy(OpStrategy) {}
   };
 
+  /// \returns true if the loop vectorizer should vectorize conditional
+  /// scalar assignments for the target.
+  bool enableCSAVectorization() const;
+
   /// \returns How the target needs this vector-predicated operation to be
   /// transformed.
   VPLegalization getVPLegalizationStrategy(const VPIntrinsic &PI) const;
@@ -2275,6 +2279,7 @@ class TargetTransformInfo::Concept {
                              SmallVectorImpl<Use *> &OpsToSink) const = 0;
 
   virtual bool isVectorShiftByScalarCheap(Type *Ty) const = 0;
+  virtual bool enableCSAVectorization() const = 0;
   virtual VPLegalization
   getVPLegalizationStrategy(const VPIntrinsic &PI) const = 0;
   virtual bool hasArmWideBranch(bool Thumb) const = 0;
@@ -3091,6 +3096,10 @@ class TargetTransformInfo::Model final : public TargetTransformInfo::Concept {
     return Impl.isVectorShiftByScalarCheap(Ty);
   }
 
+  bool enableCSAVectorization() const override {
+    return Impl.enableCSAVectorization();
+  }
+
   VPLegalization
   getVPLegalizationStrategy(const VPIntrinsic &PI) const override {
     return Impl.getVPLegalizationStrategy(PI);

llvm/include/llvm/Analysis/TargetTransformInfoImpl.h

@@ -1021,6 +1021,8 @@ class TargetTransformInfoImplBase {
 
   bool isVectorShiftByScalarCheap(Type *Ty) const { return false; }
 
+  bool enableCSAVectorization() const { return false; }
+
   TargetTransformInfo::VPLegalization
   getVPLegalizationStrategy(const VPIntrinsic &PI) const {
     return TargetTransformInfo::VPLegalization(

llvm/include/llvm/Transforms/Vectorize/LoopVectorizationLegality.h

@@ -269,6 +269,10 @@ class LoopVectorizationLegality {
   /// induction descriptor.
   using InductionList = MapVector<PHINode *, InductionDescriptor>;
 
+  /// CSAList contains the CSA descriptors for all the CSAs that were found
+  /// in the loop, rooted by their phis.
+  using CSAList = MapVector<PHINode *, CSADescriptor>;
+
   /// RecurrenceSet contains the phi nodes that are recurrences other than
   /// inductions and reductions.
   using RecurrenceSet = SmallPtrSet<const PHINode *, 8>;
@@ -321,6 +325,12 @@ class LoopVectorizationLegality {
   /// Returns True if V is a Phi node of an induction variable in this loop.
   bool isInductionPhi(const Value *V) const;
 
+  /// Returns the CSAs found in the loop.
+  const CSAList &getCSAs() const { return CSAs; }
+
+  /// Returns true if Phi is the root of a CSA in the loop.
+  bool isCSAPhi(PHINode *Phi) const { return CSAs.count(Phi) != 0; }
+
   /// Returns a pointer to the induction descriptor, if \p Phi is an integer or
   /// floating point induction.
   const InductionDescriptor *getIntOrFpInductionDescriptor(PHINode *Phi) const;
@@ -550,6 +560,10 @@ class LoopVectorizationLegality {
   void addInductionPhi(PHINode *Phi, const InductionDescriptor &ID,
                        SmallPtrSetImpl<Value *> &AllowedExit);
 
+  /// Updates the vetorization state by adding \p Phi to the CSA list.
+  void addCSAPhi(PHINode *Phi, const CSADescriptor &CSADesc,
+                 SmallPtrSetImpl<Value *> &AllowedExit);
+
   /// The loop that we evaluate.
   Loop *TheLoop;
 
@@ -594,6 +608,9 @@ class LoopVectorizationLegality {
   /// variables can be pointers.
   InductionList Inductions;
 
+  /// Holds the conditional scalar assignments
+  CSAList CSAs;
+
   /// Holds all the casts that participate in the update chain of the induction
   /// variables, and that have been proven to be redundant (possibly under a
   /// runtime guard). These casts can be ignored when creating the vectorized

llvm/lib/Analysis/IVDescriptors.cpp

@@ -6,7 +6,8 @@
 //
 //===----------------------------------------------------------------------===//
 //
-// This file "describes" induction and recurrence variables.
+// This file "describes" induction, recurrence, and conditional scalar
+// assignment (CSA) variables.
 //
 //===----------------------------------------------------------------------===//
 
@@ -1570,3 +1571,58 @@ bool InductionDescriptor::isInductionPHI(
   D = InductionDescriptor(StartValue, IK_PtrInduction, Step);
   return true;
 }
+
+/// Return CSADescriptor that describes a CSA that matches one of these
+/// patterns:
+///   phi loop_inv, (select cmp, value, phi)
+///   phi loop_inv, (select cmp, phi, value)
+///   phi (select cmp, value, phi), loop_inv
+///   phi (select cmp, phi, value), loop_inv
+/// If the CSA does not match any of these paterns, return a CSADescriptor
+/// that describes an InvalidCSA.
+bool CSADescriptor::isCSAPhi(PHINode *Phi, Loop *TheLoop, CSADescriptor &CSA) {
+
+  // Must be a scalar.
+  Type *Type = Phi->getType();
+  if (!Type->isIntegerTy() && !Type->isFloatingPointTy() &&
+      !Type->isPointerTy())
+    return false;
+
+  // Match phi loop_inv, (select cmp, value, phi)
+  //    or phi loop_inv, (select cmp, phi, value)
+  //    or phi (select cmp, value, phi), loop_inv
+  //    or phi (select cmp, phi, value), loop_inv
+  if (Phi->getNumIncomingValues() != 2)
+    return false;
+  auto SelectInstIt = find_if(Phi->incoming_values(), [&Phi](const Use &U) {
+    return match(U.get(), m_Select(m_Value(), m_Specific(Phi), m_Value())) ||
+           match(U.get(), m_Select(m_Value(), m_Value(), m_Specific(Phi)));
+  });
+  if (SelectInstIt == Phi->incoming_values().end())
+    return false;
+  auto LoopInvIt = find_if(Phi->incoming_values(), [&](Use &U) {
+    return U.get() != *SelectInstIt && TheLoop->isLoopInvariant(U.get());
+  });
+  if (LoopInvIt == Phi->incoming_values().end())
+    return false;
+
+  // Phi or Sel must be used only outside the loop,
+  // excluding if Phi use Sel or Sel use Phi
+  auto IsOnlyUsedOutsideLoop = [&](Value *V, Value *Ignore) {
+    return all_of(V->users(), [Ignore, TheLoop](User *U) {
+      if (U == Ignore)
+        return true;
+      if (auto *I = dyn_cast<Instruction>(U))
+        return !TheLoop->contains(I);
+      return true;
+    });
+  };
+  Instruction *Select = cast<SelectInst>(SelectInstIt->get());
+  Value *LoopInv = LoopInvIt->get();
+  if (!IsOnlyUsedOutsideLoop(Phi, Select) ||
+      !IsOnlyUsedOutsideLoop(Select, Phi))
+    return false;
+
+  CSA = CSADescriptor(Phi, Select, LoopInv);
+  return true;
+}

llvm/lib/Analysis/TargetTransformInfo.cpp

@@ -1356,6 +1356,10 @@ bool TargetTransformInfo::preferEpilogueVectorization() const {
   return TTIImpl->preferEpilogueVectorization();
 }
 
+bool TargetTransformInfo::enableCSAVectorization() const {
+  return TTIImpl->enableCSAVectorization();
+}
+
 TargetTransformInfo::VPLegalization
 TargetTransformInfo::getVPLegalizationStrategy(const VPIntrinsic &VPI) const {
   return TTIImpl->getVPLegalizationStrategy(VPI);

llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp

@@ -2370,6 +2370,11 @@ bool RISCVTTIImpl::isLegalMaskedExpandLoad(Type *DataTy, Align Alignment) {
   return true;
 }
 
+bool RISCVTTIImpl::enableCSAVectorization() const {
+  return ST->hasVInstructions() &&
+         ST->getProcFamily() == RISCVSubtarget::SiFive7;
+}
+
 bool RISCVTTIImpl::isLegalMaskedCompressStore(Type *DataTy, Align Alignment) {
   auto *VTy = dyn_cast<VectorType>(DataTy);
   if (!VTy || VTy->isScalableTy())

llvm/lib/Target/RISCV/RISCVTargetTransformInfo.h

@@ -306,6 +306,10 @@ class RISCVTTIImpl : public BasicTTIImplBase<RISCVTTIImpl> {
     return TLI->isVScaleKnownToBeAPowerOfTwo();
   }
 
+  /// \returns true if the loop vectorizer should vectorize conditional
+  /// scalar assignments for the target.
+  bool enableCSAVectorization() const;
+
   /// \returns How the target needs this vector-predicated operation to be
   /// transformed.
   TargetTransformInfo::VPLegalization

llvm/lib/Transforms/Vectorize/LoopVectorizationLegality.cpp

@@ -83,6 +83,10 @@ static cl::opt<bool> EnableHistogramVectorization(
     "enable-histogram-loop-vectorization", cl::init(false), cl::Hidden,
     cl::desc("Enables autovectorization of some loops containing histograms"));
 
+static cl::opt<bool>
+    EnableCSA("enable-csa-vectorization", cl::init(false), cl::Hidden,
+              cl::desc("Control whether CSA loop vectorization is enabled"));
+
 /// Maximum vectorization interleave count.
 static const unsigned MaxInterleaveFactor = 16;
 
@@ -749,6 +753,15 @@ bool LoopVectorizationLegality::setupOuterLoopInductions() {
   return llvm::all_of(Header->phis(), IsSupportedPhi);
 }
 
+void LoopVectorizationLegality::addCSAPhi(
+    PHINode *Phi, const CSADescriptor &CSADesc,
+    SmallPtrSetImpl<Value *> &AllowedExit) {
+  assert(CSADesc.isValid() && "Expected Valid CSADescriptor");
+  LLVM_DEBUG(dbgs() << "LV: found legal CSA opportunity" << *Phi << "\n");
+  AllowedExit.insert(Phi);
+  CSAs.insert({Phi, CSADesc});
+}
+
 /// Checks if a function is scalarizable according to the TLI, in
 /// the sense that it should be vectorized and then expanded in
 /// multiple scalar calls. This is represented in the
@@ -866,14 +879,24 @@ bool LoopVectorizationLegality::canVectorizeInstrs() {
           continue;
         }
 
-        // As a last resort, coerce the PHI to a AddRec expression
-        // and re-try classifying it a an induction PHI.
+        // Try to coerce the PHI to a AddRec expression and re-try classifying
+        // it a an induction PHI.
         if (InductionDescriptor::isInductionPHI(Phi, TheLoop, PSE, ID, true) &&
             !IsDisallowedStridedPointerInduction(ID)) {
           addInductionPhi(Phi, ID, AllowedExit);
           continue;
         }
 
+        // Check if the PHI can be classified as a CSA PHI.
+        if (EnableCSA || (TTI->enableCSAVectorization() &&
+                          EnableCSA.getNumOccurrences() == 0)) {
+          CSADescriptor CSADesc;
+          if (CSADescriptor::isCSAPhi(Phi, TheLoop, CSADesc)) {
+            addCSAPhi(Phi, CSADesc, AllowedExit);
+            continue;
+          }
+        }
+
         reportVectorizationFailure("Found an unidentified PHI",
             "value that could not be identified as "
             "reduction is used outside the loop",
@@ -1844,11 +1867,15 @@ bool LoopVectorizationLegality::canFoldTailByMasking() const {
   for (const auto &Reduction : getReductionVars())
     ReductionLiveOuts.insert(Reduction.second.getLoopExitInstr());
 
+  SmallPtrSet<const Value *, 8> CSALiveOuts;
+  for (const auto &CSA : getCSAs())
+    CSALiveOuts.insert(CSA.second.getAssignment());
+
   // TODO: handle non-reduction outside users when tail is folded by masking.
   for (auto *AE : AllowedExit) {
     // Check that all users of allowed exit values are inside the loop or
-    // are the live-out of a reduction.
-    if (ReductionLiveOuts.count(AE))
+    // are the live-out of a reduction or CSA.
+    if (ReductionLiveOuts.count(AE) || CSALiveOuts.count(AE))
       continue;
     for (User *U : AE->users()) {
       Instruction *UI = cast<Instruction>(U);

llvm/lib/Transforms/Vectorize/LoopVectorizationPlanner.h

@@ -174,8 +174,8 @@ class VPBuilder {
         new VPInstruction(Opcode, Operands, WrapFlags, DL, Name));
   }
 
-  VPValue *createNot(VPValue *Operand, DebugLoc DL = {},
-                     const Twine &Name = "") {
+  VPInstruction *createNot(VPValue *Operand, DebugLoc DL = {},
+                           const Twine &Name = "") {
     return createInstruction(VPInstruction::Not, {Operand}, DL, Name);
   }
 
@@ -261,6 +261,22 @@ class VPBuilder {
         FPBinOp ? FPBinOp->getFastMathFlags() : FastMathFlags()));
   }
 
+  VPInstruction *createCSAMaskPhi(VPValue *InitMask, DebugLoc DL,
+                                  const Twine &Name) {
+    return createInstruction(VPInstruction::CSAMaskPhi, {InitMask}, DL, Name);
+  }
+
+  VPInstruction *createAnyOf(VPValue *Cond, DebugLoc DL, const Twine &Name) {
+    return createInstruction(VPInstruction::AnyOf, {Cond}, DL, Name);
+  }
+
+  VPInstruction *createCSAMaskSel(VPValue *Cond, VPValue *MaskPhi,
+                                  VPValue *AnyOf, DebugLoc DL,
+                                  const Twine &Name) {
+    return createInstruction(VPInstruction::CSAMaskSel, {Cond, MaskPhi, AnyOf},
+                             DL, Name);
+  }
+
   //===--------------------------------------------------------------------===//
   // RAII helpers.
   //===--------------------------------------------------------------------===//