[llvm] [VPlan] Create header phis once, after constructing VPlan0 (NFC). (PR #168291)
Florian Hahn via llvm-commits
llvm-commits at lists.llvm.org
Thu Nov 20 12:52:37 PST 2025
https://github.com/fhahn updated https://github.com/llvm/llvm-project/pull/168291
>From 506c53b52644e8d0a8c88db34791347c0b0f5749 Mon Sep 17 00:00:00 2001
From: Florian Hahn <flo at fhahn.com>
Date: Sun, 2 Nov 2025 21:10:19 +0000
Subject: [PATCH 1/2] [VPlan] Create resume phis in scalar preheader early.
(NFC)
Create phi recipes for scalar resume value up front in
addInitialSkeleton during initial construction. This will allow moving
the remaining code dealing with resume values to VPlan
transforms/construction.
---
llvm/lib/Transforms/Vectorize/VPlan.h | 4 ++-
.../Vectorize/VPlanConstruction.cpp | 9 +++++
.../Transforms/Vectorize/VPlanTransforms.cpp | 36 ++++++++-----------
.../Transforms/Vectorize/VPlanHCFGTest.cpp | 6 ++--
.../Vectorize/VPlanVerifierTest.cpp | 4 +--
5 files changed, 33 insertions(+), 26 deletions(-)
diff --git a/llvm/lib/Transforms/Vectorize/VPlan.h b/llvm/lib/Transforms/Vectorize/VPlan.h
index 0fc527e07e634..a15584cd0c6ea 100644
--- a/llvm/lib/Transforms/Vectorize/VPlan.h
+++ b/llvm/lib/Transforms/Vectorize/VPlan.h
@@ -1134,7 +1134,7 @@ class LLVM_ABI_FOR_TEST VPInstruction : public VPRecipeWithIRFlags,
OpcodeTy Opcode;
/// An optional name that can be used for the generated IR instruction.
- const std::string Name;
+ std::string Name;
/// Returns true if we can generate a scalar for the first lane only if
/// needed.
@@ -1225,6 +1225,8 @@ class LLVM_ABI_FOR_TEST VPInstruction : public VPRecipeWithIRFlags,
/// Returns the symbolic name assigned to the VPInstruction.
StringRef getName() const { return Name; }
+ void setName(StringRef NewName) { Name = NewName.str(); }
+
protected:
#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
/// Print the VPInstruction to \p O.
diff --git a/llvm/lib/Transforms/Vectorize/VPlanConstruction.cpp b/llvm/lib/Transforms/Vectorize/VPlanConstruction.cpp
index dbbde1cafa9f2..5fbd61a929fe2 100644
--- a/llvm/lib/Transforms/Vectorize/VPlanConstruction.cpp
+++ b/llvm/lib/Transforms/Vectorize/VPlanConstruction.cpp
@@ -554,6 +554,15 @@ static void addInitialSkeleton(VPlan &Plan, Type *InductionTy, DebugLoc IVDL,
Plan.getEntry()->swapSuccessors();
createExtractsForLiveOuts(Plan, MiddleVPBB);
+
+ VPBuilder ScalarPHBuilder(ScalarPH);
+ for (const auto &[PhiR, ScalarPhiR] : zip_equal(
+ drop_begin(HeaderVPBB->phis()), Plan.getScalarHeader()->phis())) {
+ auto *VectorPhiR = cast<VPPhi>(&PhiR);
+ auto *ResumePhiR = ScalarPHBuilder.createScalarPhi(
+ {VectorPhiR, VectorPhiR->getOperand(0)}, VectorPhiR->getDebugLoc());
+ cast<VPIRPhi>(&ScalarPhiR)->addOperand(ResumePhiR);
+ }
}
std::unique_ptr<VPlan>
diff --git a/llvm/lib/Transforms/Vectorize/VPlanTransforms.cpp b/llvm/lib/Transforms/Vectorize/VPlanTransforms.cpp
index 25557f1d5d651..916e8b1cea28e 100644
--- a/llvm/lib/Transforms/Vectorize/VPlanTransforms.cpp
+++ b/llvm/lib/Transforms/Vectorize/VPlanTransforms.cpp
@@ -4527,9 +4527,10 @@ void VPlanTransforms::addBranchWeightToMiddleTerminator(
/// Create and return a ResumePhi for \p WideIV, unless it is truncated. If the
/// induction recipe is not canonical, creates a VPDerivedIVRecipe to compute
/// the end value of the induction.
-static VPInstruction *addResumePhiRecipeForInduction(
- VPWidenInductionRecipe *WideIV, VPBuilder &VectorPHBuilder,
- VPBuilder &ScalarPHBuilder, VPTypeAnalysis &TypeInfo, VPValue *VectorTC) {
+static VPValue *addResumePhiRecipeForInduction(VPWidenInductionRecipe *WideIV,
+ VPBuilder &VectorPHBuilder,
+ VPTypeAnalysis &TypeInfo,
+ VPValue *VectorTC) {
auto *WideIntOrFp = dyn_cast<VPWidenIntOrFpInductionRecipe>(WideIV);
// Truncated wide inductions resume from the last lane of their vector value
// in the last vector iteration which is handled elsewhere.
@@ -4555,9 +4556,7 @@ static VPInstruction *addResumePhiRecipeForInduction(
WideIV->getDebugLoc());
}
- auto *ResumePhiRecipe = ScalarPHBuilder.createScalarPhi(
- {EndValue, Start}, WideIV->getDebugLoc(), "bc.resume.val");
- return ResumePhiRecipe;
+ return EndValue;
}
void VPlanTransforms::addScalarResumePhis(
@@ -4570,21 +4569,18 @@ void VPlanTransforms::addScalarResumePhis(
VPBuilder VectorPHBuilder(
cast<VPBasicBlock>(VectorRegion->getSinglePredecessor()));
VPBuilder MiddleBuilder(MiddleVPBB, MiddleVPBB->getFirstNonPhi());
- VPBuilder ScalarPHBuilder(ScalarPH);
- for (VPRecipeBase &ScalarPhiR : Plan.getScalarHeader()->phis()) {
- auto *ScalarPhiIRI = cast<VPIRPhi>(&ScalarPhiR);
+ for (VPRecipeBase &PhiR : Plan.getScalarPreheader()->phis()) {
+ auto *ResumePhiR = cast<VPPhi>(&PhiR);
// TODO: Extract final value from induction recipe initially, optimize to
// pre-computed end value together in optimizeInductionExitUsers.
- auto *VectorPhiR =
- cast<VPHeaderPHIRecipe>(Builder.getRecipe(&ScalarPhiIRI->getIRPhi()));
+ auto *VectorPhiR = cast<VPHeaderPHIRecipe>(ResumePhiR->getOperand(0));
if (auto *WideIVR = dyn_cast<VPWidenInductionRecipe>(VectorPhiR)) {
- if (VPInstruction *ResumePhi = addResumePhiRecipeForInduction(
- WideIVR, VectorPHBuilder, ScalarPHBuilder, TypeInfo,
- &Plan.getVectorTripCount())) {
- assert(isa<VPPhi>(ResumePhi) && "Expected a phi");
- IVEndValues[WideIVR] = ResumePhi->getOperand(0);
- ScalarPhiIRI->addOperand(ResumePhi);
+ if (VPValue *ResumeV = addResumePhiRecipeForInduction(
+ WideIVR, VectorPHBuilder, TypeInfo, &Plan.getVectorTripCount())) {
+ IVEndValues[WideIVR] = ResumeV;
+ ResumePhiR->setOperand(0, ResumeV);
+ ResumePhiR->setName("bc.resume.val");
continue;
}
// TODO: Also handle truncated inductions here. Computing end-values
@@ -4606,10 +4602,8 @@ void VPlanTransforms::addScalarResumePhis(
ResumeFromVectorLoop = MiddleBuilder.createNaryOp(
VPInstruction::ExtractLastElement, {ResumeFromVectorLoop}, {},
"vector.recur.extract");
- StringRef Name = IsFOR ? "scalar.recur.init" : "bc.merge.rdx";
- auto *ResumePhiR = ScalarPHBuilder.createScalarPhi(
- {ResumeFromVectorLoop, VectorPhiR->getStartValue()}, {}, Name);
- ScalarPhiIRI->addOperand(ResumePhiR);
+ ResumePhiR->setName(IsFOR ? "scalar.recur.init" : "bc.merge.rdx");
+ ResumePhiR->setOperand(0, ResumeFromVectorLoop);
}
}
diff --git a/llvm/unittests/Transforms/Vectorize/VPlanHCFGTest.cpp b/llvm/unittests/Transforms/Vectorize/VPlanHCFGTest.cpp
index 5742df2aa3c53..127e830b1bc14 100644
--- a/llvm/unittests/Transforms/Vectorize/VPlanHCFGTest.cpp
+++ b/llvm/unittests/Transforms/Vectorize/VPlanHCFGTest.cpp
@@ -113,12 +113,13 @@ compound=true
N0 -> N2 [ label="F"]
N1 [label =
"scalar.ph:\l" +
+ " EMIT-SCALAR vp\<%6\> = phi [ ir\<%indvars.iv\>, middle.block ], [ ir\<0\>, ir-bb\<entry\> ]\l" +
"Successor(s): ir-bb\<for.body\>\l"
]
N1 -> N3 [ label=""]
N3 [label =
"ir-bb\<for.body\>:\l" +
- " IR %indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ]\l" +
+ " IR %indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ] (extra operand: vp\<%6\> from scalar.ph)\l" +
" IR %arr.idx = getelementptr inbounds i32, ptr %A, i64 %indvars.iv\l" +
" IR %l1 = load i32, ptr %arr.idx, align 4\l" +
" IR %res = add i32 %l1, 10\l" +
@@ -282,12 +283,13 @@ compound=true
N0 -> N2 [ label="F"]
N1 [label =
"scalar.ph:\l" +
+ " EMIT-SCALAR vp\<%6\> = phi [ ir\<%iv\>, middle.block ], [ ir\<0\>, ir-bb\<entry\> ]\l" +
"Successor(s): ir-bb\<loop.header\>\l"
]
N1 -> N3 [ label=""]
N3 [label =
"ir-bb\<loop.header\>:\l" +
- " IR %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop.latch ]\l" +
+ " IR %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop.latch ] (extra operand: vp\<%6\> from scalar.ph)\l" +
" IR %arr.idx = getelementptr inbounds i32, ptr %A, i64 %iv\l" +
" IR %l1 = load i32, ptr %arr.idx, align 4\l" +
" IR %c = icmp eq i32 %l1, 0\l" +
diff --git a/llvm/unittests/Transforms/Vectorize/VPlanVerifierTest.cpp b/llvm/unittests/Transforms/Vectorize/VPlanVerifierTest.cpp
index 1b0cd1074c1b5..ec8f06a1eb755 100644
--- a/llvm/unittests/Transforms/Vectorize/VPlanVerifierTest.cpp
+++ b/llvm/unittests/Transforms/Vectorize/VPlanVerifierTest.cpp
@@ -346,6 +346,8 @@ TEST_F(VPIRVerifierTest, testVerifyIRPhiInScalarHeaderVPIRBB) {
Function *F = M.getFunction("f");
BasicBlock *LoopHeader = F->getEntryBlock().getSingleSuccessor();
auto Plan = buildVPlan(LoopHeader);
+ VPValue *Zero = Plan->getConstantInt(32, 0);
+ Plan->getScalarHeader()->front().addOperand(Zero);
#if GTEST_HAS_STREAM_REDIRECTION
::testing::internal::CaptureStderr();
@@ -387,8 +389,6 @@ TEST_F(VPIRVerifierTest, testVerifyIRPhiInExitVPIRBB) {
{HeaderBlock->front().getVPSingleValue()});
DefI->insertBefore(Plan->getMiddleBlock()->getTerminator());
Plan->getExitBlocks()[0]->front().addOperand(DefI);
- VPValue *Zero = Plan->getConstantInt(32, 0);
- Plan->getScalarHeader()->front().addOperand(Zero);
#if GTEST_HAS_STREAM_REDIRECTION
::testing::internal::CaptureStderr();
>From a086412f55bcacfc1fc0c8f369814a3dd06a500a Mon Sep 17 00:00:00 2001
From: Florian Hahn <flo at fhahn.com>
Date: Sun, 16 Nov 2025 21:37:24 +0000
Subject: [PATCH 2/2] [VPlan] Create header phis once, after constructing
VPlan0 (NFC).
---
.../Transforms/Vectorize/LoopVectorize.cpp | 122 +++---------------
.../Transforms/Vectorize/VPRecipeBuilder.h | 4 -
llvm/lib/Transforms/Vectorize/VPlan.h | 11 +-
.../Vectorize/VPlanConstruction.cpp | 92 +++++++++++++
.../Transforms/Vectorize/VPlanTransforms.h | 11 ++
5 files changed, 133 insertions(+), 107 deletions(-)
diff --git a/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp b/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
index 5b8725178521e..ed05e3d6c712c 100644
--- a/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
+++ b/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
@@ -1412,6 +1412,11 @@ class LoopVectorizationCostModel {
return InLoopReductions.contains(Phi);
}
+ /// Returns the set of in-loop reduction PHIs.
+ const SmallPtrSetImpl<PHINode *> &getInLoopReductions() const {
+ return InLoopReductions;
+ }
+
/// Returns true if the predicated reduction select should be used to set the
/// incoming value for the reduction phi.
bool usePredicatedReductionSelect() const {
@@ -7627,57 +7632,6 @@ VPWidenMemoryRecipe *VPRecipeBuilder::tryToWidenMemory(VPInstruction *VPI,
Consecutive, Reverse, *VPI, VPI->getDebugLoc());
}
-/// Creates a VPWidenIntOrFpInductionRecipe for \p PhiR. If needed, it will
-/// also insert a recipe to expand the step for the induction recipe.
-static VPWidenIntOrFpInductionRecipe *
-createWidenInductionRecipes(VPInstruction *PhiR,
- const InductionDescriptor &IndDesc, VPlan &Plan,
- ScalarEvolution &SE, Loop &OrigLoop) {
- assert(SE.isLoopInvariant(IndDesc.getStep(), &OrigLoop) &&
- "step must be loop invariant");
-
- VPValue *Start = PhiR->getOperand(0);
- assert(Plan.getLiveIn(IndDesc.getStartValue()) == Start &&
- "Start VPValue must match IndDesc's start value");
-
- // It is always safe to copy over the NoWrap and FastMath flags. In
- // particular, when folding tail by masking, the masked-off lanes are never
- // used, so it is safe.
- VPIRFlags Flags = vputils::getFlagsFromIndDesc(IndDesc);
- VPValue *Step =
- vputils::getOrCreateVPValueForSCEVExpr(Plan, IndDesc.getStep());
-
- // Update wide induction increments to use the same step as the corresponding
- // wide induction. This enables detecting induction increments directly in
- // VPlan and removes redundant splats.
- using namespace llvm::VPlanPatternMatch;
- if (match(PhiR->getOperand(1), m_Add(m_Specific(PhiR), m_VPValue())))
- PhiR->getOperand(1)->getDefiningRecipe()->setOperand(1, Step);
-
- PHINode *Phi = cast<PHINode>(PhiR->getUnderlyingInstr());
- return new VPWidenIntOrFpInductionRecipe(Phi, Start, Step, &Plan.getVF(),
- IndDesc, Flags, PhiR->getDebugLoc());
-}
-
-VPHeaderPHIRecipe *
-VPRecipeBuilder::tryToOptimizeInductionPHI(VPInstruction *VPI) {
- auto *Phi = cast<PHINode>(VPI->getUnderlyingInstr());
-
- // Check if this is an integer or fp induction. If so, build the recipe that
- // produces its scalar and vector values.
- if (auto *II = Legal->getIntOrFpInductionDescriptor(Phi))
- return createWidenInductionRecipes(VPI, *II, Plan, *PSE.getSE(), *OrigLoop);
-
- // Check if this is pointer induction. If so, build the recipe for it.
- if (auto *II = Legal->getPointerInductionDescriptor(Phi)) {
- VPValue *Step = vputils::getOrCreateVPValueForSCEVExpr(Plan, II->getStep());
- return new VPWidenPointerInductionRecipe(Phi, VPI->getOperand(0), Step,
- &Plan.getVFxUF(), *II,
- VPI->getDebugLoc());
- }
- return nullptr;
-}
-
VPWidenIntOrFpInductionRecipe *
VPRecipeBuilder::tryToOptimizeInductionTruncate(VPInstruction *VPI,
VFRange &Range) {
@@ -8186,45 +8140,7 @@ VPRecipeBase *VPRecipeBuilder::tryToCreateWidenRecipe(VPSingleDefRecipe *R,
// First, check for specific widening recipes that deal with inductions, Phi
// nodes, calls and memory operations.
VPRecipeBase *Recipe;
- if (auto *PhiR = dyn_cast<VPPhi>(R)) {
- VPBasicBlock *Parent = PhiR->getParent();
- [[maybe_unused]] VPRegionBlock *LoopRegionOf =
- Parent->getEnclosingLoopRegion();
- assert(LoopRegionOf && LoopRegionOf->getEntry() == Parent &&
- "Non-header phis should have been handled during predication");
- auto *Phi = cast<PHINode>(R->getUnderlyingInstr());
- assert(R->getNumOperands() == 2 && "Must have 2 operands for header phis");
- if ((Recipe = tryToOptimizeInductionPHI(PhiR)))
- return Recipe;
-
- VPHeaderPHIRecipe *PhiRecipe = nullptr;
- assert((Legal->isReductionVariable(Phi) ||
- Legal->isFixedOrderRecurrence(Phi)) &&
- "can only widen reductions and fixed-order recurrences here");
- VPValue *StartV = R->getOperand(0);
- if (Legal->isReductionVariable(Phi)) {
- const RecurrenceDescriptor &RdxDesc = Legal->getRecurrenceDescriptor(Phi);
- assert(RdxDesc.getRecurrenceStartValue() ==
- Phi->getIncomingValueForBlock(OrigLoop->getLoopPreheader()));
-
- // If the PHI is used by a partial reduction, set the scale factor.
- unsigned ScaleFactor =
- getScalingForReduction(RdxDesc.getLoopExitInstr()).value_or(1);
- PhiRecipe = new VPReductionPHIRecipe(
- Phi, RdxDesc.getRecurrenceKind(), *StartV, CM.isInLoopReduction(Phi),
- CM.useOrderedReductions(RdxDesc), ScaleFactor);
- } else {
- // TODO: Currently fixed-order recurrences are modeled as chains of
- // first-order recurrences. If there are no users of the intermediate
- // recurrences in the chain, the fixed order recurrence should be modeled
- // directly, enabling more efficient codegen.
- PhiRecipe = new VPFirstOrderRecurrencePHIRecipe(Phi, *StartV);
- }
- // Add backedge value.
- PhiRecipe->addOperand(R->getOperand(1));
- return PhiRecipe;
- }
- assert(!R->isPhi() && "only VPPhi nodes expected at this point");
+ assert(!R->isPhi() && "phis must be handled earlier");
auto *VPI = cast<VPInstruction>(R);
Instruction *Instr = R->getUnderlyingInstr();
@@ -8284,6 +8200,9 @@ VPRecipeBuilder::tryToCreatePartialReduction(VPInstruction *Reduction,
if (isa<VPReductionPHIRecipe>(BinOp) || isa<VPPartialReductionRecipe>(BinOp))
std::swap(BinOp, Accumulator);
+ if (auto *RedPhiR = dyn_cast<VPReductionPHIRecipe>(Accumulator))
+ RedPhiR->setVFScaleFactor(ScaleFactor);
+
assert(ScaleFactor ==
vputils::getVFScaleFactor(Accumulator->getDefiningRecipe()) &&
"all accumulators in chain must have same scale factor");
@@ -8331,6 +8250,12 @@ void LoopVectorizationPlanner::buildVPlansWithVPRecipes(ElementCount MinVF,
OrigLoop, *LI, Legal->getWidestInductionType(),
getDebugLocFromInstOrOperands(Legal->getPrimaryInduction()), PSE, &LVer);
+ // Create recipes for header phis.
+ VPlanTransforms::createHeaderPhiRecipes(
+ *VPlan0, *PSE.getSE(), *OrigLoop, Legal->getInductionVars(),
+ Legal->getReductionVars(), Legal->getFixedOrderRecurrences(),
+ CM.getInLoopReductions(), Hints.allowReordering());
+
auto MaxVFTimes2 = MaxVF * 2;
for (ElementCount VF = MinVF; ElementCount::isKnownLT(VF, MaxVFTimes2);) {
VFRange SubRange = {VF, MaxVFTimes2};
@@ -8451,25 +8376,18 @@ VPlanPtr LoopVectorizationPlanner::tryToBuildVPlanWithVPRecipes(
// Mapping from VPValues in the initial plan to their widened VPValues. Needed
// temporarily to update created block masks.
DenseMap<VPValue *, VPValue *> Old2New;
+
+ // Now process all other blocks and instructions.
for (VPBasicBlock *VPBB : VPBlockUtils::blocksOnly<VPBasicBlock>(RPOT)) {
// Convert input VPInstructions to widened recipes.
for (VPRecipeBase &R : make_early_inc_range(*VPBB)) {
- auto *SingleDef = cast<VPSingleDefRecipe>(&R);
- auto *UnderlyingValue = SingleDef->getUnderlyingValue();
- // Skip recipes that do not need transforming, including canonical IV,
- // wide canonical IV and VPInstructions without underlying values. The
- // latter are added above for masking.
- // FIXME: Migrate code relying on the underlying instruction from VPlan0
- // to construct recipes below to not use the underlying instruction.
- if (isa<VPCanonicalIVPHIRecipe, VPWidenCanonicalIVRecipe, VPBlendRecipe>(
- &R) ||
- (isa<VPInstruction>(&R) && !UnderlyingValue))
+ auto *SingleDef = dyn_cast<VPInstruction>(&R);
+ if (!SingleDef || !SingleDef->getUnderlyingValue())
continue;
- assert(isa<VPInstruction>(&R) && UnderlyingValue && "unsupported recipe");
// TODO: Gradually replace uses of underlying instruction by analyses on
// VPlan.
- Instruction *Instr = cast<Instruction>(UnderlyingValue);
+ Instruction *Instr = cast<Instruction>(SingleDef->getUnderlyingValue());
Builder.setInsertPoint(SingleDef);
// The stores with invariant address inside the loop will be deleted, and
diff --git a/llvm/lib/Transforms/Vectorize/VPRecipeBuilder.h b/llvm/lib/Transforms/Vectorize/VPRecipeBuilder.h
index 1808be118cd2a..fb96d0c213f3b 100644
--- a/llvm/lib/Transforms/Vectorize/VPRecipeBuilder.h
+++ b/llvm/lib/Transforms/Vectorize/VPRecipeBuilder.h
@@ -94,10 +94,6 @@ class VPRecipeBuilder {
/// recipe that takes an additional VPInstruction for the mask.
VPWidenMemoryRecipe *tryToWidenMemory(VPInstruction *VPI, VFRange &Range);
- /// Check if an induction recipe should be constructed for \p VPI. If so build
- /// and return it. If not, return null.
- VPHeaderPHIRecipe *tryToOptimizeInductionPHI(VPInstruction *VPI);
-
/// Optimize the special case where the operand of \p VPI is a constant
/// integer induction variable.
VPWidenIntOrFpInductionRecipe *
diff --git a/llvm/lib/Transforms/Vectorize/VPlan.h b/llvm/lib/Transforms/Vectorize/VPlan.h
index a15584cd0c6ea..4f6a67c012ac9 100644
--- a/llvm/lib/Transforms/Vectorize/VPlan.h
+++ b/llvm/lib/Transforms/Vectorize/VPlan.h
@@ -2377,8 +2377,10 @@ struct VPFirstOrderRecurrencePHIRecipe : public VPHeaderPHIRecipe {
VP_CLASSOF_IMPL(VPDef::VPFirstOrderRecurrencePHISC)
VPFirstOrderRecurrencePHIRecipe *clone() override {
- return new VPFirstOrderRecurrencePHIRecipe(
+ auto *R = new VPFirstOrderRecurrencePHIRecipe(
cast<PHINode>(getUnderlyingInstr()), *getOperand(0));
+ R->addOperand(getOperand(1));
+ return R;
}
void execute(VPTransformState &State) override;
@@ -2448,6 +2450,13 @@ class VPReductionPHIRecipe : public VPHeaderPHIRecipe,
/// Get the factor that the VF of this recipe's output should be scaled by.
unsigned getVFScaleFactor() const { return VFScaleFactor; }
+ /// Set the VFScaleFactor for this reduction phi. Can only be set to a factor
+ /// > 1
+ void setVFScaleFactor(unsigned ScaleFactor) {
+ assert(ScaleFactor > 1 && "must set to scale factor > 1");
+ VFScaleFactor = ScaleFactor;
+ }
+
/// Returns the number of incoming values, also number of incoming blocks.
/// Note that at the moment, VPWidenPointerInductionRecipe only has a single
/// incoming value, its start value.
diff --git a/llvm/lib/Transforms/Vectorize/VPlanConstruction.cpp b/llvm/lib/Transforms/Vectorize/VPlanConstruction.cpp
index 5fbd61a929fe2..348f95d0dd56a 100644
--- a/llvm/lib/Transforms/Vectorize/VPlanConstruction.cpp
+++ b/llvm/lib/Transforms/Vectorize/VPlanConstruction.cpp
@@ -575,6 +575,98 @@ VPlanTransforms::buildVPlan0(Loop *TheLoop, LoopInfo &LI, Type *InductionTy,
return VPlan0;
}
+/// Creates a VPWidenIntOrFpInductionRecipe or VPWidenPointerInductionRecipe
+/// for \p Phi based on \p IndDesc.
+static VPHeaderPHIRecipe *
+createWidenInductionRecipe(VPPhi *PhiR, const InductionDescriptor &IndDesc,
+ VPlan &Plan, ScalarEvolution &SE, Loop &OrigLoop) {
+ assert(SE.isLoopInvariant(IndDesc.getStep(), &OrigLoop) &&
+ "step must be loop invariant");
+
+ VPValue *Start = PhiR->getOperand(0);
+ assert(Plan.getLiveIn(IndDesc.getStartValue()) == Start &&
+ "Start VPValue must match IndDesc's start value");
+ VPValue *Step =
+ vputils::getOrCreateVPValueForSCEVExpr(Plan, IndDesc.getStep());
+
+ if (IndDesc.getKind() == InductionDescriptor::IK_PtrInduction)
+ return new VPWidenPointerInductionRecipe(
+ cast<PHINode>(PhiR->getUnderlyingInstr()), Start, Step,
+ &Plan.getVFxUF(), IndDesc, PhiR->getDebugLoc());
+ // Update wide induction increments to use the same step as the corresponding
+ // wide induction. This enables detecting induction increments directly in
+ // VPlan and removes redundant splats.
+ using namespace llvm::VPlanPatternMatch;
+ if (match(PhiR->getOperand(1), m_Add(m_Specific(PhiR), m_VPValue())))
+ PhiR->getOperand(1)->getDefiningRecipe()->setOperand(1, Step);
+
+ // It is always safe to copy over the NoWrap and FastMath flags. In
+ // particular, when folding tail by masking, the masked-off lanes are never
+ // used, so it is safe.
+ VPIRFlags Flags = vputils::getFlagsFromIndDesc(IndDesc);
+
+ return new VPWidenIntOrFpInductionRecipe(
+ cast<PHINode>(PhiR->getUnderlyingInstr()), Start, Step, &Plan.getVF(),
+ IndDesc, Flags, PhiR->getDebugLoc());
+}
+
+void VPlanTransforms::createHeaderPhiRecipes(
+ VPlan &Plan, ScalarEvolution &SE, Loop &OrigLoop,
+ const MapVector<PHINode *, InductionDescriptor> &Inductions,
+ const MapVector<PHINode *, RecurrenceDescriptor> &Reductions,
+ const SmallPtrSetImpl<const PHINode *> &FixedOrderRecurrences,
+ const SmallPtrSetImpl<PHINode *> &InLoopReductions, bool AllowReordering) {
+
+ VPBasicBlock *HeaderVPBB = cast<VPBasicBlock>(
+ Plan.getEntry()->getSuccessors()[1]->getSingleSuccessor());
+
+ for (VPRecipeBase &R : make_early_inc_range(*HeaderVPBB)) {
+ if (isa<VPCanonicalIVPHIRecipe>(&R))
+ continue;
+ auto *PhiR = dyn_cast<VPPhi>(&R);
+ if (!PhiR)
+ break;
+
+ // TODO: Gradually replace uses of underlying instruction by analyses on
+ // VPlan.
+ auto *Phi = cast<PHINode>(PhiR->getUnderlyingInstr());
+ assert(PhiR->getNumOperands() == 2 &&
+ "Must have 2 operands for header phis");
+
+ VPHeaderPHIRecipe *HeaderPhiR = nullptr;
+ auto InductionIt = Inductions.find(Phi);
+ if (InductionIt != Inductions.end()) {
+ HeaderPhiR = createWidenInductionRecipe(PhiR, InductionIt->second, Plan,
+ SE, OrigLoop);
+ } else {
+ VPValue *Start = PhiR->getOperand(0);
+ auto ReductionIt = Reductions.find(Phi);
+ if (ReductionIt != Reductions.end()) {
+ const RecurrenceDescriptor &RdxDesc = ReductionIt->second;
+ assert(RdxDesc.getRecurrenceStartValue() ==
+ Phi->getIncomingValueForBlock(OrigLoop.getLoopPreheader()));
+
+ bool UseOrderedReductions = !AllowReordering && RdxDesc.isOrdered();
+ HeaderPhiR = new VPReductionPHIRecipe(
+ Phi, RdxDesc.getRecurrenceKind(), *Start,
+ InLoopReductions.contains(Phi), UseOrderedReductions);
+ } else {
+ assert(FixedOrderRecurrences.contains(Phi) &&
+ "can only widen reductions and fixed-order recurrences here");
+ // TODO: Currently fixed-order recurrences are modeled as chains of
+ // first-order recurrences. If there are no users of the intermediate
+ // recurrences in the chain, the fixed order recurrence should be
+ // modeled directly, enabling more efficient codegen.
+ HeaderPhiR = new VPFirstOrderRecurrencePHIRecipe(Phi, *Start);
+ }
+ HeaderPhiR->addOperand(PhiR->getOperand(1));
+ }
+ HeaderPhiR->insertBefore(PhiR);
+ PhiR->replaceAllUsesWith(HeaderPhiR);
+ PhiR->eraseFromParent();
+ }
+}
+
void VPlanTransforms::handleEarlyExits(VPlan &Plan,
bool HasUncountableEarlyExit) {
auto *MiddleVPBB = cast<VPBasicBlock>(
diff --git a/llvm/lib/Transforms/Vectorize/VPlanTransforms.h b/llvm/lib/Transforms/Vectorize/VPlanTransforms.h
index 708ea4185e1cb..fc13ea5696eb2 100644
--- a/llvm/lib/Transforms/Vectorize/VPlanTransforms.h
+++ b/llvm/lib/Transforms/Vectorize/VPlanTransforms.h
@@ -102,6 +102,17 @@ struct VPlanTransforms {
buildVPlan0(Loop *TheLoop, LoopInfo &LI, Type *InductionTy, DebugLoc IVDL,
PredicatedScalarEvolution &PSE, LoopVersioning *LVer = nullptr);
+ /// Replace VPPhi recipes in \p Plan's header with corresponding
+ /// VPHeaderPHIRecipe subclasses for inductions, reductions, and
+ /// fixed-order recurrences. This processes all header phis and creates
+ /// the appropriate widened recipe for each one.
+ static void createHeaderPhiRecipes(
+ VPlan &Plan, ScalarEvolution &SE, Loop &OrigLoop,
+ const MapVector<PHINode *, InductionDescriptor> &Inductions,
+ const MapVector<PHINode *, RecurrenceDescriptor> &Reductions,
+ const SmallPtrSetImpl<const PHINode *> &FixedOrderRecurrences,
+ const SmallPtrSetImpl<PHINode *> &InLoopReductions, bool AllowReordering);
+
/// Update \p Plan to account for all early exits.
LLVM_ABI_FOR_TEST static void handleEarlyExits(VPlan &Plan,
bool HasUncountableExit);
More information about the llvm-commits
mailing list