blob: 1a6065cb3f1a6cb1039da4c48c86578473c08acf [file] [log] [blame]
//===- LoopUnroll.cpp - Loop unroller pass --------------------------------===//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
// This pass implements a simple loop unroller. It works best when loops have
// been canonicalized by the -indvars pass, allowing it to determine the trip
// counts of loops easily.
#include "llvm/Transforms/Scalar/LoopUnrollPass.h"
#include "llvm/ADT/DenseMap.h"
#include "llvm/ADT/DenseMapInfo.h"
#include "llvm/ADT/DenseSet.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/ADT/SetVector.h"
#include "llvm/ADT/SmallPtrSet.h"
#include "llvm/ADT/SmallVector.h"
#include "llvm/ADT/StringRef.h"
#include "llvm/Analysis/AssumptionCache.h"
#include "llvm/Analysis/BlockFrequencyInfo.h"
#include "llvm/Analysis/CodeMetrics.h"
#include "llvm/Analysis/LoopAnalysisManager.h"
#include "llvm/Analysis/LoopInfo.h"
#include "llvm/Analysis/LoopPass.h"
#include "llvm/Analysis/LoopUnrollAnalyzer.h"
#include "llvm/Analysis/OptimizationRemarkEmitter.h"
#include "llvm/Analysis/ProfileSummaryInfo.h"
#include "llvm/Analysis/ScalarEvolution.h"
#include "llvm/Analysis/TargetTransformInfo.h"
#include "llvm/IR/BasicBlock.h"
#include "llvm/IR/CFG.h"
#include "llvm/IR/Constant.h"
#include "llvm/IR/Constants.h"
#include "llvm/IR/DiagnosticInfo.h"
#include "llvm/IR/Dominators.h"
#include "llvm/IR/Function.h"
#include "llvm/IR/Instruction.h"
#include "llvm/IR/Instructions.h"
#include "llvm/IR/IntrinsicInst.h"
#include "llvm/IR/Metadata.h"
#include "llvm/IR/PassManager.h"
#include "llvm/InitializePasses.h"
#include "llvm/Pass.h"
#include "llvm/Support/Casting.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/ErrorHandling.h"
#include "llvm/Support/raw_ostream.h"
#include "llvm/Transforms/Scalar.h"
#include "llvm/Transforms/Scalar/LoopPassManager.h"
#include "llvm/Transforms/Utils.h"
#include "llvm/Transforms/Utils/LoopPeel.h"
#include "llvm/Transforms/Utils/LoopSimplify.h"
#include "llvm/Transforms/Utils/LoopUtils.h"
#include "llvm/Transforms/Utils/SizeOpts.h"
#include "llvm/Transforms/Utils/UnrollLoop.h"
#include <algorithm>
#include <cassert>
#include <cstdint>
#include <limits>
#include <optional>
#include <string>
#include <tuple>
#include <utility>
using namespace llvm;
#define DEBUG_TYPE "loop-unroll"
cl::opt<bool> llvm::ForgetSCEVInLoopUnroll(
"forget-scev-loop-unroll", cl::init(false), cl::Hidden,
cl::desc("Forget everything in SCEV when doing LoopUnroll, instead of just"
" the current top-most loop. This is sometimes preferred to reduce"
" compile time."));
static cl::opt<unsigned>
UnrollThreshold("unroll-threshold", cl::Hidden,
cl::desc("The cost threshold for loop unrolling"));
static cl::opt<unsigned>
"unroll-optsize-threshold", cl::init(0), cl::Hidden,
cl::desc("The cost threshold for loop unrolling when optimizing for "
static cl::opt<unsigned> UnrollPartialThreshold(
"unroll-partial-threshold", cl::Hidden,
cl::desc("The cost threshold for partial loop unrolling"));
static cl::opt<unsigned> UnrollMaxPercentThresholdBoost(
"unroll-max-percent-threshold-boost", cl::init(400), cl::Hidden,
cl::desc("The maximum 'boost' (represented as a percentage >= 100) applied "
"to the threshold when aggressively unrolling a loop due to the "
"dynamic cost savings. If completely unrolling a loop will reduce "
"the total runtime from X to Y, we boost the loop unroll "
"threshold to DefaultThreshold*std::min(MaxPercentThresholdBoost, "
"X/Y). This limit avoids excessive code bloat."));
static cl::opt<unsigned> UnrollMaxIterationsCountToAnalyze(
"unroll-max-iteration-count-to-analyze", cl::init(10), cl::Hidden,
cl::desc("Don't allow loop unrolling to simulate more than this number of"
"iterations when checking full unroll profitability"));
static cl::opt<unsigned> UnrollCount(
"unroll-count", cl::Hidden,
cl::desc("Use this unroll count for all loops including those with "
"unroll_count pragma values, for testing purposes"));
static cl::opt<unsigned> UnrollMaxCount(
"unroll-max-count", cl::Hidden,
cl::desc("Set the max unroll count for partial and runtime unrolling, for"
"testing purposes"));
static cl::opt<unsigned> UnrollFullMaxCount(
"unroll-full-max-count", cl::Hidden,
"Set the max unroll count for full unrolling, for testing purposes"));
static cl::opt<bool>
UnrollAllowPartial("unroll-allow-partial", cl::Hidden,
cl::desc("Allows loops to be partially unrolled until "
"-unroll-threshold loop size is reached."));
static cl::opt<bool> UnrollAllowRemainder(
"unroll-allow-remainder", cl::Hidden,
cl::desc("Allow generation of a loop remainder (extra iterations) "
"when unrolling a loop."));
static cl::opt<bool>
UnrollRuntime("unroll-runtime", cl::Hidden,
cl::desc("Unroll loops with run-time trip counts"));
static cl::opt<unsigned> UnrollMaxUpperBound(
"unroll-max-upperbound", cl::init(8), cl::Hidden,
"The max of trip count upper bound that is considered in unrolling"));
static cl::opt<unsigned> PragmaUnrollThreshold(
"pragma-unroll-threshold", cl::init(16 * 1024), cl::Hidden,
cl::desc("Unrolled size limit for loops with an unroll(full) or "
"unroll_count pragma."));
static cl::opt<unsigned> FlatLoopTripCountThreshold(
"flat-loop-tripcount-threshold", cl::init(5), cl::Hidden,
cl::desc("If the runtime tripcount for the loop is lower than the "
"threshold, the loop is considered as flat and will be less "
"aggressively unrolled."));
static cl::opt<bool> UnrollUnrollRemainder(
"unroll-remainder", cl::Hidden,
cl::desc("Allow the loop remainder to be unrolled."));
// This option isn't ever intended to be enabled, it serves to allow
// experiments to check the assumptions about when this kind of revisit is
// necessary.
static cl::opt<bool> UnrollRevisitChildLoops(
"unroll-revisit-child-loops", cl::Hidden,
cl::desc("Enqueue and re-visit child loops in the loop PM after unrolling. "
"This shouldn't typically be needed as child loops (or their "
"clones) were already visited."));
static cl::opt<unsigned> UnrollThresholdAggressive(
"unroll-threshold-aggressive", cl::init(300), cl::Hidden,
cl::desc("Threshold (max size of unrolled loop) to use in aggressive (O3) "
static cl::opt<unsigned>
UnrollThresholdDefault("unroll-threshold-default", cl::init(150),
cl::desc("Default threshold (max size of unrolled "
"loop), used in all but O3 optimizations"));
/// A magic value for use with the Threshold parameter to indicate
/// that the loop unroll should be performed regardless of how much
/// code expansion would result.
static const unsigned NoThreshold = std::numeric_limits<unsigned>::max();
/// Gather the various unrolling parameters based on the defaults, compiler
/// flags, TTI overrides and user specified parameters.
TargetTransformInfo::UnrollingPreferences llvm::gatherUnrollingPreferences(
Loop *L, ScalarEvolution &SE, const TargetTransformInfo &TTI,
BlockFrequencyInfo *BFI, ProfileSummaryInfo *PSI,
OptimizationRemarkEmitter &ORE, int OptLevel,
std::optional<unsigned> UserThreshold, std::optional<unsigned> UserCount,
std::optional<bool> UserAllowPartial, std::optional<bool> UserRuntime,
std::optional<bool> UserUpperBound,
std::optional<unsigned> UserFullUnrollMaxCount) {
TargetTransformInfo::UnrollingPreferences UP;
// Set up the defaults
UP.Threshold =
OptLevel > 2 ? UnrollThresholdAggressive : UnrollThresholdDefault;
UP.MaxPercentThresholdBoost = 400;
UP.OptSizeThreshold = UnrollOptSizeThreshold;
UP.PartialThreshold = 150;
UP.PartialOptSizeThreshold = UnrollOptSizeThreshold;
UP.Count = 0;
UP.DefaultUnrollRuntimeCount = 8;
UP.MaxCount = std::numeric_limits<unsigned>::max();
UP.FullUnrollMaxCount = std::numeric_limits<unsigned>::max();
UP.BEInsns = 2;
UP.Partial = false;
UP.Runtime = false;
UP.AllowRemainder = true;
UP.UnrollRemainder = false;
UP.AllowExpensiveTripCount = false;
UP.Force = false;
UP.UpperBound = false;
UP.UnrollAndJam = false;
UP.UnrollAndJamInnerLoopThreshold = 60;
UP.MaxIterationsCountToAnalyze = UnrollMaxIterationsCountToAnalyze;
// Override with any target specific settings
TTI.getUnrollingPreferences(L, SE, UP, &ORE);
// Apply size attributes
bool OptForSize = L->getHeader()->getParent()->hasOptSize() ||
// Let unroll hints / pragmas take precedence over PGSO.
(hasUnrollTransformation(L) != TM_ForcedByUser &&
llvm::shouldOptimizeForSize(L->getHeader(), PSI, BFI,
if (OptForSize) {
UP.Threshold = UP.OptSizeThreshold;
UP.PartialThreshold = UP.PartialOptSizeThreshold;
UP.MaxPercentThresholdBoost = 100;
// Apply any user values specified by cl::opt
if (UnrollThreshold.getNumOccurrences() > 0)
UP.Threshold = UnrollThreshold;
if (UnrollPartialThreshold.getNumOccurrences() > 0)
UP.PartialThreshold = UnrollPartialThreshold;
if (UnrollMaxPercentThresholdBoost.getNumOccurrences() > 0)
UP.MaxPercentThresholdBoost = UnrollMaxPercentThresholdBoost;
if (UnrollMaxCount.getNumOccurrences() > 0)
UP.MaxCount = UnrollMaxCount;
if (UnrollFullMaxCount.getNumOccurrences() > 0)
UP.FullUnrollMaxCount = UnrollFullMaxCount;
if (UnrollAllowPartial.getNumOccurrences() > 0)
UP.Partial = UnrollAllowPartial;
if (UnrollAllowRemainder.getNumOccurrences() > 0)
UP.AllowRemainder = UnrollAllowRemainder;
if (UnrollRuntime.getNumOccurrences() > 0)
UP.Runtime = UnrollRuntime;
if (UnrollMaxUpperBound == 0)
UP.UpperBound = false;
if (UnrollUnrollRemainder.getNumOccurrences() > 0)
UP.UnrollRemainder = UnrollUnrollRemainder;
if (UnrollMaxIterationsCountToAnalyze.getNumOccurrences() > 0)
UP.MaxIterationsCountToAnalyze = UnrollMaxIterationsCountToAnalyze;
// Apply user values provided by argument
if (UserThreshold) {
UP.Threshold = *UserThreshold;
UP.PartialThreshold = *UserThreshold;
if (UserCount)
UP.Count = *UserCount;
if (UserAllowPartial)
UP.Partial = *UserAllowPartial;
if (UserRuntime)
UP.Runtime = *UserRuntime;
if (UserUpperBound)
UP.UpperBound = *UserUpperBound;
if (UserFullUnrollMaxCount)
UP.FullUnrollMaxCount = *UserFullUnrollMaxCount;
return UP;
namespace {
/// A struct to densely store the state of an instruction after unrolling at
/// each iteration.
/// This is designed to work like a tuple of <Instruction *, int> for the
/// purposes of hashing and lookup, but to be able to associate two boolean
/// states with each key.
struct UnrolledInstState {
Instruction *I;
int Iteration : 30;
unsigned IsFree : 1;
unsigned IsCounted : 1;
/// Hashing and equality testing for a set of the instruction states.
struct UnrolledInstStateKeyInfo {
using PtrInfo = DenseMapInfo<Instruction *>;
using PairInfo = DenseMapInfo<std::pair<Instruction *, int>>;
static inline UnrolledInstState getEmptyKey() {
return {PtrInfo::getEmptyKey(), 0, 0, 0};
static inline UnrolledInstState getTombstoneKey() {
return {PtrInfo::getTombstoneKey(), 0, 0, 0};
static inline unsigned getHashValue(const UnrolledInstState &S) {
return PairInfo::getHashValue({S.I, S.Iteration});
static inline bool isEqual(const UnrolledInstState &LHS,
const UnrolledInstState &RHS) {
return PairInfo::isEqual({LHS.I, LHS.Iteration}, {RHS.I, RHS.Iteration});
struct EstimatedUnrollCost {
/// The estimated cost after unrolling.
unsigned UnrolledCost;
/// The estimated dynamic cost of executing the instructions in the
/// rolled form.
unsigned RolledDynamicCost;
struct PragmaInfo {
PragmaInfo(bool UUC, bool PFU, unsigned PC, bool PEU)
: UserUnrollCount(UUC), PragmaFullUnroll(PFU), PragmaCount(PC),
PragmaEnableUnroll(PEU) {}
const bool UserUnrollCount;
const bool PragmaFullUnroll;
const unsigned PragmaCount;
const bool PragmaEnableUnroll;
} // end anonymous namespace
/// Figure out if the loop is worth full unrolling.
/// Complete loop unrolling can make some loads constant, and we need to know
/// if that would expose any further optimization opportunities. This routine
/// estimates this optimization. It computes cost of unrolled loop
/// (UnrolledCost) and dynamic cost of the original loop (RolledDynamicCost). By
/// dynamic cost we mean that we won't count costs of blocks that are known not
/// to be executed (i.e. if we have a branch in the loop and we know that at the
/// given iteration its condition would be resolved to true, we won't add up the
/// cost of the 'false'-block).
/// \returns Optional value, holding the RolledDynamicCost and UnrolledCost. If
/// the analysis failed (no benefits expected from the unrolling, or the loop is
/// too big to analyze), the returned value is std::nullopt.
static std::optional<EstimatedUnrollCost> analyzeLoopUnrollCost(
const Loop *L, unsigned TripCount, DominatorTree &DT, ScalarEvolution &SE,
const SmallPtrSetImpl<const Value *> &EphValues,
const TargetTransformInfo &TTI, unsigned MaxUnrolledLoopSize,
unsigned MaxIterationsCountToAnalyze) {
// We want to be able to scale offsets by the trip count and add more offsets
// to them without checking for overflows, and we already don't want to
// analyze *massive* trip counts, so we force the max to be reasonably small.
assert(MaxIterationsCountToAnalyze <
(unsigned)(std::numeric_limits<int>::max() / 2) &&
"The unroll iterations max is too large!");
// Only analyze inner loops. We can't properly estimate cost of nested loops
// and we won't visit inner loops again anyway.
if (!L->isInnermost())
return std::nullopt;
// Don't simulate loops with a big or unknown tripcount
if (!TripCount || TripCount > MaxIterationsCountToAnalyze)
return std::nullopt;
SmallSetVector<BasicBlock *, 16> BBWorklist;
SmallSetVector<std::pair<BasicBlock *, BasicBlock *>, 4> ExitWorklist;
DenseMap<Value *, Value *> SimplifiedValues;
SmallVector<std::pair<Value *, Value *>, 4> SimplifiedInputValues;
// The estimated cost of the unrolled form of the loop. We try to estimate
// this by simplifying as much as we can while computing the estimate.
InstructionCost UnrolledCost = 0;
// We also track the estimated dynamic (that is, actually executed) cost in
// the rolled form. This helps identify cases when the savings from unrolling
// aren't just exposing dead control flows, but actual reduced dynamic
// instructions due to the simplifications which we expect to occur after
// unrolling.
InstructionCost RolledDynamicCost = 0;
// We track the simplification of each instruction in each iteration. We use
// this to recursively merge costs into the unrolled cost on-demand so that
// we don't count the cost of any dead code. This is essentially a map from
// <instruction, int> to <bool, bool>, but stored as a densely packed struct.
DenseSet<UnrolledInstState, UnrolledInstStateKeyInfo> InstCostMap;
// A small worklist used to accumulate cost of instructions from each
// observable and reached root in the loop.
SmallVector<Instruction *, 16> CostWorklist;
// PHI-used worklist used between iterations while accumulating cost.
SmallVector<Instruction *, 4> PHIUsedList;
// Helper function to accumulate cost for instructions in the loop.
auto AddCostRecursively = [&](Instruction &RootI, int Iteration) {
assert(Iteration >= 0 && "Cannot have a negative iteration!");
assert(CostWorklist.empty() && "Must start with an empty cost list");
assert(PHIUsedList.empty() && "Must start with an empty phi used list");
TargetTransformInfo::TargetCostKind CostKind =
RootI.getFunction()->hasMinSize() ?
TargetTransformInfo::TCK_CodeSize :
for (;; --Iteration) {
do {
Instruction *I = CostWorklist.pop_back_val();
// InstCostMap only uses I and Iteration as a key, the other two values
// don't matter here.
auto CostIter = InstCostMap.find({I, Iteration, 0, 0});
if (CostIter == InstCostMap.end())
// If an input to a PHI node comes from a dead path through the loop
// we may have no cost data for it here. What that actually means is
// that it is free.
auto &Cost = *CostIter;
if (Cost.IsCounted)
// Already counted this instruction.
// Mark that we are counting the cost of this instruction now.
Cost.IsCounted = true;
// If this is a PHI node in the loop header, just add it to the PHI set.
if (auto *PhiI = dyn_cast<PHINode>(I))
if (PhiI->getParent() == L->getHeader()) {
assert(Cost.IsFree && "Loop PHIs shouldn't be evaluated as they "
"inherently simplify during unrolling.");
if (Iteration == 0)
// Push the incoming value from the backedge into the PHI used list
// if it is an in-loop instruction. We'll use this to populate the
// cost worklist for the next iteration (as we count backwards).
if (auto *OpI = dyn_cast<Instruction>(
if (L->contains(OpI))
// First accumulate the cost of this instruction.
if (!Cost.IsFree) {
UnrolledCost += TTI.getInstructionCost(I, CostKind);
LLVM_DEBUG(dbgs() << "Adding cost of instruction (iteration "
<< Iteration << "): ");
// We must count the cost of every operand which is not free,
// recursively. If we reach a loop PHI node, simply add it to the set
// to be considered on the next iteration (backwards!).
for (Value *Op : I->operands()) {
// Check whether this operand is free due to being a constant or
// outside the loop.
auto *OpI = dyn_cast<Instruction>(Op);
if (!OpI || !L->contains(OpI))
// Otherwise accumulate its cost.
} while (!CostWorklist.empty());
if (PHIUsedList.empty())
// We've exhausted the search.
assert(Iteration > 0 &&
"Cannot track PHI-used values past the first iteration!");
CostWorklist.append(PHIUsedList.begin(), PHIUsedList.end());
// Ensure that we don't violate the loop structure invariants relied on by
// this analysis.
assert(L->isLoopSimplifyForm() && "Must put loop into normal form first.");
assert(L->isLCSSAForm(DT) &&
"Must have loops in LCSSA form to track live-out values.");
LLVM_DEBUG(dbgs() << "Starting LoopUnroll profitability analysis...\n");
TargetTransformInfo::TargetCostKind CostKind =
L->getHeader()->getParent()->hasMinSize() ?
TargetTransformInfo::TCK_CodeSize : TargetTransformInfo::TCK_SizeAndLatency;
// Simulate execution of each iteration of the loop counting instructions,
// which would be simplified.
// Since the same load will take different values on different iterations,
// we literally have to go through all loop's iterations.
for (unsigned Iteration = 0; Iteration < TripCount; ++Iteration) {
LLVM_DEBUG(dbgs() << " Analyzing iteration " << Iteration << "\n");
// Prepare for the iteration by collecting any simplified entry or backedge
// inputs.
for (Instruction &I : *L->getHeader()) {
auto *PHI = dyn_cast<PHINode>(&I);
if (!PHI)
// The loop header PHI nodes must have exactly two input: one from the
// loop preheader and one from the loop latch.
PHI->getNumIncomingValues() == 2 &&
"Must have an incoming value only for the preheader and the latch.");
Value *V = PHI->getIncomingValueForBlock(
Iteration == 0 ? L->getLoopPreheader() : L->getLoopLatch());
if (Iteration != 0 && SimplifiedValues.count(V))
V = SimplifiedValues.lookup(V);
SimplifiedInputValues.push_back({PHI, V});
// Now clear and re-populate the map for the next iteration.
while (!SimplifiedInputValues.empty())
UnrolledInstAnalyzer Analyzer(Iteration, SimplifiedValues, SE, L);
// Note that we *must not* cache the size, this loop grows the worklist.
for (unsigned Idx = 0; Idx != BBWorklist.size(); ++Idx) {
BasicBlock *BB = BBWorklist[Idx];
// Visit all instructions in the given basic block and try to simplify
// it. We don't change the actual IR, just count optimization
// opportunities.
for (Instruction &I : *BB) {
// These won't get into the final code - don't even try calculating the
// cost for them.
if (isa<DbgInfoIntrinsic>(I) || EphValues.count(&I))
// Track this instruction's expected baseline cost when executing the
// rolled loop form.
RolledDynamicCost += TTI.getInstructionCost(&I, CostKind);
// Visit the instruction to analyze its loop cost after unrolling,
// and if the visitor returns true, mark the instruction as free after
// unrolling and continue.
bool IsFree = Analyzer.visit(I);
bool Inserted = InstCostMap.insert({&I, (int)Iteration,
/*IsCounted*/ false}).second;
assert(Inserted && "Cannot have a state for an unvisited instruction!");
if (IsFree)
// Can't properly model a cost of a call.
// FIXME: With a proper cost model we should be able to do it.
if (auto *CI = dyn_cast<CallInst>(&I)) {
const Function *Callee = CI->getCalledFunction();
if (!Callee || TTI.isLoweredToCall(Callee)) {
LLVM_DEBUG(dbgs() << "Can't analyze cost of loop with call\n");
return std::nullopt;
// If the instruction might have a side-effect recursively account for
// the cost of it and all the instructions leading up to it.
if (I.mayHaveSideEffects())
AddCostRecursively(I, Iteration);
// If unrolled body turns out to be too big, bail out.
if (UnrolledCost > MaxUnrolledLoopSize) {
LLVM_DEBUG(dbgs() << " Exceeded threshold.. exiting.\n"
<< " UnrolledCost: " << UnrolledCost
<< ", MaxUnrolledLoopSize: " << MaxUnrolledLoopSize
<< "\n");
return std::nullopt;
Instruction *TI = BB->getTerminator();
auto getSimplifiedConstant = [&](Value *V) -> Constant * {
if (SimplifiedValues.count(V))
V = SimplifiedValues.lookup(V);
return dyn_cast<Constant>(V);
// Add in the live successors by first checking whether we have terminator
// that may be simplified based on the values simplified by this call.
BasicBlock *KnownSucc = nullptr;
if (BranchInst *BI = dyn_cast<BranchInst>(TI)) {
if (BI->isConditional()) {
if (auto *SimpleCond = getSimplifiedConstant(BI->getCondition())) {
// Just take the first successor if condition is undef
if (isa<UndefValue>(SimpleCond))
KnownSucc = BI->getSuccessor(0);
else if (ConstantInt *SimpleCondVal =
KnownSucc = BI->getSuccessor(SimpleCondVal->isZero() ? 1 : 0);
} else if (SwitchInst *SI = dyn_cast<SwitchInst>(TI)) {
if (auto *SimpleCond = getSimplifiedConstant(SI->getCondition())) {
// Just take the first successor if condition is undef
if (isa<UndefValue>(SimpleCond))
KnownSucc = SI->getSuccessor(0);
else if (ConstantInt *SimpleCondVal =
KnownSucc = SI->findCaseValue(SimpleCondVal)->getCaseSuccessor();
if (KnownSucc) {
if (L->contains(KnownSucc))
ExitWorklist.insert({BB, KnownSucc});
// Add BB's successors to the worklist.
for (BasicBlock *Succ : successors(BB))
if (L->contains(Succ))
ExitWorklist.insert({BB, Succ});
AddCostRecursively(*TI, Iteration);
// If we found no optimization opportunities on the first iteration, we
// won't find them on later ones too.
if (UnrolledCost == RolledDynamicCost) {
LLVM_DEBUG(dbgs() << " No opportunities found.. exiting.\n"
<< " UnrolledCost: " << UnrolledCost << "\n");
return std::nullopt;
while (!ExitWorklist.empty()) {
BasicBlock *ExitingBB, *ExitBB;
std::tie(ExitingBB, ExitBB) = ExitWorklist.pop_back_val();
for (Instruction &I : *ExitBB) {
auto *PN = dyn_cast<PHINode>(&I);
if (!PN)
Value *Op = PN->getIncomingValueForBlock(ExitingBB);
if (auto *OpI = dyn_cast<Instruction>(Op))
if (L->contains(OpI))
AddCostRecursively(*OpI, TripCount - 1);
assert(UnrolledCost.isValid() && RolledDynamicCost.isValid() &&
"All instructions must have a valid cost, whether the "
"loop is rolled or unrolled.");
LLVM_DEBUG(dbgs() << "Analysis finished:\n"
<< "UnrolledCost: " << UnrolledCost << ", "
<< "RolledDynamicCost: " << RolledDynamicCost << "\n");
return {{unsigned(*UnrolledCost.getValue()),
/// ApproximateLoopSize - Approximate the size of the loop.
InstructionCost llvm::ApproximateLoopSize(
const Loop *L, unsigned &NumCalls, bool &NotDuplicatable, bool &Convergent,
const TargetTransformInfo &TTI,
const SmallPtrSetImpl<const Value *> &EphValues, unsigned BEInsns) {
CodeMetrics Metrics;
for (BasicBlock *BB : L->blocks())
Metrics.analyzeBasicBlock(BB, TTI, EphValues);
NumCalls = Metrics.NumInlineCandidates;
NotDuplicatable = Metrics.notDuplicatable;
Convergent = Metrics.convergent;
InstructionCost LoopSize = Metrics.NumInsts;
// Don't allow an estimate of size zero. This would allows unrolling of loops
// with huge iteration counts, which is a compile time problem even if it's
// not a problem for code quality. Also, the code using this size may assume
// that each loop has at least three instructions (likely a conditional
// branch, a comparison feeding that branch, and some kind of loop increment
// feeding that comparison instruction).
if (LoopSize.isValid() && LoopSize < BEInsns + 1)
// This is an open coded max() on InstructionCost
LoopSize = BEInsns + 1;
return LoopSize;
// Returns the loop hint metadata node with the given name (for example,
// "llvm.loop.unroll.count"). If no such metadata node exists, then nullptr is
// returned.
static MDNode *getUnrollMetadataForLoop(const Loop *L, StringRef Name) {
if (MDNode *LoopID = L->getLoopID())
return GetUnrollMetadata(LoopID, Name);
return nullptr;
// Returns true if the loop has an unroll(full) pragma.
static bool hasUnrollFullPragma(const Loop *L) {
return getUnrollMetadataForLoop(L, "llvm.loop.unroll.full");
// Returns true if the loop has an unroll(enable) pragma. This metadata is used
// for both "#pragma unroll" and "#pragma clang loop unroll(enable)" directives.
static bool hasUnrollEnablePragma(const Loop *L) {
return getUnrollMetadataForLoop(L, "llvm.loop.unroll.enable");
// Returns true if the loop has an runtime unroll(disable) pragma.
static bool hasRuntimeUnrollDisablePragma(const Loop *L) {
return getUnrollMetadataForLoop(L, "llvm.loop.unroll.runtime.disable");
// If loop has an unroll_count pragma return the (necessarily
// positive) value from the pragma. Otherwise return 0.
static unsigned unrollCountPragmaValue(const Loop *L) {
MDNode *MD = getUnrollMetadataForLoop(L, "llvm.loop.unroll.count");
if (MD) {
assert(MD->getNumOperands() == 2 &&
"Unroll count hint metadata should have two operands.");
unsigned Count =
assert(Count >= 1 && "Unroll count must be positive.");
return Count;
return 0;
// Computes the boosting factor for complete unrolling.
// If fully unrolling the loop would save a lot of RolledDynamicCost, it would
// be beneficial to fully unroll the loop even if unrolledcost is large. We
// use (RolledDynamicCost / UnrolledCost) to model the unroll benefits to adjust
// the unroll threshold.
static unsigned getFullUnrollBoostingFactor(const EstimatedUnrollCost &Cost,
unsigned MaxPercentThresholdBoost) {
if (Cost.RolledDynamicCost >= std::numeric_limits<unsigned>::max() / 100)
return 100;
else if (Cost.UnrolledCost != 0)
// The boosting factor is RolledDynamicCost / UnrolledCost
return std::min(100 * Cost.RolledDynamicCost / Cost.UnrolledCost,
return MaxPercentThresholdBoost;
// Produce an estimate of the unrolled cost of the specified loop. This
// is used to a) produce a cost estimate for partial unrolling and b) to
// cheaply estimate cost for full unrolling when we don't want to symbolically
// evaluate all iterations.
class UnrollCostEstimator {
const unsigned LoopSize;
UnrollCostEstimator(Loop &L, unsigned LoopSize) : LoopSize(LoopSize) {}
// Returns loop size estimation for unrolled loop, given the unrolling
// configuration specified by UP.
getUnrolledLoopSize(const TargetTransformInfo::UnrollingPreferences &UP,
const unsigned CountOverwrite = 0) const {
assert(LoopSize >= UP.BEInsns &&
"LoopSize should not be less than BEInsns!");
if (CountOverwrite)
return static_cast<uint64_t>(LoopSize - UP.BEInsns) * CountOverwrite +
return static_cast<uint64_t>(LoopSize - UP.BEInsns) * UP.Count +
static std::optional<unsigned>
shouldPragmaUnroll(Loop *L, const PragmaInfo &PInfo,
const unsigned TripMultiple, const unsigned TripCount,
const UnrollCostEstimator UCE,
const TargetTransformInfo::UnrollingPreferences &UP) {
// Using unroll pragma
// 1st priority is unroll count set by "unroll-count" option.
if (PInfo.UserUnrollCount) {
if (UP.AllowRemainder &&
UCE.getUnrolledLoopSize(UP, (unsigned)UnrollCount) < UP.Threshold)
return (unsigned)UnrollCount;
// 2nd priority is unroll count set by pragma.
if (PInfo.PragmaCount > 0) {
if ((UP.AllowRemainder || (TripMultiple % PInfo.PragmaCount == 0)))
return PInfo.PragmaCount;
if (PInfo.PragmaFullUnroll && TripCount != 0)
return TripCount;
// if didn't return until here, should continue to other priorties
return std::nullopt;
static std::optional<unsigned> shouldFullUnroll(
Loop *L, const TargetTransformInfo &TTI, DominatorTree &DT,
ScalarEvolution &SE, const SmallPtrSetImpl<const Value *> &EphValues,
const unsigned FullUnrollTripCount, const UnrollCostEstimator UCE,
const TargetTransformInfo::UnrollingPreferences &UP) {
assert(FullUnrollTripCount && "should be non-zero!");
if (FullUnrollTripCount > UP.FullUnrollMaxCount)
return std::nullopt;
// When computing the unrolled size, note that BEInsns are not replicated
// like the rest of the loop body.
if (UCE.getUnrolledLoopSize(UP) < UP.Threshold)
return FullUnrollTripCount;
// The loop isn't that small, but we still can fully unroll it if that
// helps to remove a significant number of instructions.
// To check that, run additional analysis on the loop.
if (std::optional<EstimatedUnrollCost> Cost = analyzeLoopUnrollCost(
L, FullUnrollTripCount, DT, SE, EphValues, TTI,
UP.Threshold * UP.MaxPercentThresholdBoost / 100,
UP.MaxIterationsCountToAnalyze)) {
unsigned Boost =
getFullUnrollBoostingFactor(*Cost, UP.MaxPercentThresholdBoost);
if (Cost->UnrolledCost < UP.Threshold * Boost / 100)
return FullUnrollTripCount;
return std::nullopt;
static std::optional<unsigned>
shouldPartialUnroll(const unsigned LoopSize, const unsigned TripCount,
const UnrollCostEstimator UCE,
const TargetTransformInfo::UnrollingPreferences &UP) {
if (!TripCount)
return std::nullopt;
if (!UP.Partial) {
LLVM_DEBUG(dbgs() << " will not try to unroll partially because "
<< "-unroll-allow-partial not given\n");
return 0;
unsigned count = UP.Count;
if (count == 0)
count = TripCount;
if (UP.PartialThreshold != NoThreshold) {
// Reduce unroll count to be modulo of TripCount for partial unrolling.
if (UCE.getUnrolledLoopSize(UP, count) > UP.PartialThreshold)
count = (std::max(UP.PartialThreshold, UP.BEInsns + 1) - UP.BEInsns) /
(LoopSize - UP.BEInsns);
if (count > UP.MaxCount)
count = UP.MaxCount;
while (count != 0 && TripCount % count != 0)
if (UP.AllowRemainder && count <= 1) {
// If there is no Count that is modulo of TripCount, set Count to
// largest power-of-two factor that satisfies the threshold limit.
// As we'll create fixup loop, do the type of unrolling only if
// remainder loop is allowed.
count = UP.DefaultUnrollRuntimeCount;
while (count != 0 &&
UCE.getUnrolledLoopSize(UP, count) > UP.PartialThreshold)
count >>= 1;
if (count < 2) {
count = 0;
} else {
count = TripCount;
if (count > UP.MaxCount)
count = UP.MaxCount;
LLVM_DEBUG(dbgs() << " partially unrolling with count: " << count << "\n");
return count;
// Returns true if unroll count was set explicitly.
// Calculates unroll count and writes it to UP.Count.
// Unless IgnoreUser is true, will also use metadata and command-line options
// that are specific to to the LoopUnroll pass (which, for instance, are
// irrelevant for the LoopUnrollAndJam pass).
// FIXME: This function is used by LoopUnroll and LoopUnrollAndJam, but consumes
// many LoopUnroll-specific options. The shared functionality should be
// refactored into it own function.
bool llvm::computeUnrollCount(
Loop *L, const TargetTransformInfo &TTI, DominatorTree &DT, LoopInfo *LI,
AssumptionCache *AC,
ScalarEvolution &SE, const SmallPtrSetImpl<const Value *> &EphValues,
OptimizationRemarkEmitter *ORE, unsigned TripCount, unsigned MaxTripCount,
bool MaxOrZero, unsigned TripMultiple, unsigned LoopSize,
TargetTransformInfo::UnrollingPreferences &UP,
TargetTransformInfo::PeelingPreferences &PP, bool &UseUpperBound) {
UnrollCostEstimator UCE(*L, LoopSize);
const bool UserUnrollCount = UnrollCount.getNumOccurrences() > 0;
const bool PragmaFullUnroll = hasUnrollFullPragma(L);
const unsigned PragmaCount = unrollCountPragmaValue(L);
const bool PragmaEnableUnroll = hasUnrollEnablePragma(L);
const bool ExplicitUnroll = PragmaCount > 0 || PragmaFullUnroll ||
PragmaEnableUnroll || UserUnrollCount;
PragmaInfo PInfo(UserUnrollCount, PragmaFullUnroll, PragmaCount,
// Use an explicit peel count that has been specified for testing. In this
// case it's not permitted to also specify an explicit unroll count.
if (PP.PeelCount) {
if (UnrollCount.getNumOccurrences() > 0) {
report_fatal_error("Cannot specify both explicit peel count and "
"explicit unroll count", /*GenCrashDiag=*/false);
UP.Count = 1;
UP.Runtime = false;
return true;
// Check for explicit Count.
// 1st priority is unroll count set by "unroll-count" option.
// 2nd priority is unroll count set by pragma.
if (auto UnrollFactor = shouldPragmaUnroll(L, PInfo, TripMultiple, TripCount,
UCE, UP)) {
UP.Count = *UnrollFactor;
if (UserUnrollCount || (PragmaCount > 0)) {
UP.AllowExpensiveTripCount = true;
UP.Force = true;
UP.Runtime |= (PragmaCount > 0);
return ExplicitUnroll;
} else {
if (ExplicitUnroll && TripCount != 0) {
// If the loop has an unrolling pragma, we want to be more aggressive with
// unrolling limits. Set thresholds to at least the PragmaUnrollThreshold
// value which is larger than the default limits.
UP.Threshold = std::max<unsigned>(UP.Threshold, PragmaUnrollThreshold);
UP.PartialThreshold =
std::max<unsigned>(UP.PartialThreshold, PragmaUnrollThreshold);
// 3rd priority is exact full unrolling. This will eliminate all copies
// of some exit test.
UP.Count = 0;
if (TripCount) {
UP.Count = TripCount;
if (auto UnrollFactor = shouldFullUnroll(L, TTI, DT, SE, EphValues,
TripCount, UCE, UP)) {
UP.Count = *UnrollFactor;
UseUpperBound = false;
return ExplicitUnroll;
// 4th priority is bounded unrolling.
// We can unroll by the upper bound amount if it's generally allowed or if
// we know that the loop is executed either the upper bound or zero times.
// (MaxOrZero unrolling keeps only the first loop test, so the number of
// loop tests remains the same compared to the non-unrolled version, whereas
// the generic upper bound unrolling keeps all but the last loop test so the
// number of loop tests goes up which may end up being worse on targets with
// constrained branch predictor resources so is controlled by an option.)
// In addition we only unroll small upper bounds.
// Note that the cost of bounded unrolling is always strictly greater than
// cost of exact full unrolling. As such, if we have an exact count and
// found it unprofitable, we'll never chose to bounded unroll.
if (!TripCount && MaxTripCount && (UP.UpperBound || MaxOrZero) &&
MaxTripCount <= UnrollMaxUpperBound) {
UP.Count = MaxTripCount;
if (auto UnrollFactor = shouldFullUnroll(L, TTI, DT, SE, EphValues,
MaxTripCount, UCE, UP)) {
UP.Count = *UnrollFactor;
UseUpperBound = true;
return ExplicitUnroll;
// 5th priority is loop peeling.
computePeelCount(L, LoopSize, PP, TripCount, DT, SE, AC, UP.Threshold);
if (PP.PeelCount) {
UP.Runtime = false;
UP.Count = 1;
return ExplicitUnroll;
// Before starting partial unrolling, set up.partial to true,
// if user explicitly asked for unrolling
if (TripCount)
UP.Partial |= ExplicitUnroll;
// 6th priority is partial unrolling.
// Try partial unroll only when TripCount could be statically calculated.
if (auto UnrollFactor = shouldPartialUnroll(LoopSize, TripCount, UCE, UP)) {
UP.Count = *UnrollFactor;
if ((PragmaFullUnroll || PragmaEnableUnroll) && TripCount &&
UP.Count != TripCount)
ORE->emit([&]() {
return OptimizationRemarkMissed(DEBUG_TYPE,
L->getStartLoc(), L->getHeader())
<< "Unable to fully unroll loop as directed by unroll pragma "
"because "
"unrolled size is too large.";
if (UP.PartialThreshold != NoThreshold) {
if (UP.Count == 0) {
if (PragmaEnableUnroll)
ORE->emit([&]() {
return OptimizationRemarkMissed(DEBUG_TYPE,
L->getStartLoc(), L->getHeader())
<< "Unable to unroll loop as directed by unroll(enable) "
"pragma "
"because unrolled size is too large.";
return ExplicitUnroll;
assert(TripCount == 0 &&
"All cases when TripCount is constant should be covered here.");
if (PragmaFullUnroll)
ORE->emit([&]() {
return OptimizationRemarkMissed(
DEBUG_TYPE, "CantFullUnrollAsDirectedRuntimeTripCount",
L->getStartLoc(), L->getHeader())
<< "Unable to fully unroll loop as directed by unroll(full) "
"pragma "
"because loop has a runtime trip count.";
// 7th priority is runtime unrolling.
// Don't unroll a runtime trip count loop when it is disabled.
if (hasRuntimeUnrollDisablePragma(L)) {
UP.Count = 0;
return false;
// Don't unroll a small upper bound loop unless user or TTI asked to do so.
if (MaxTripCount && !UP.Force && MaxTripCount < UnrollMaxUpperBound) {
UP.Count = 0;
return false;
// Check if the runtime trip count is too small when profile is available.
if (L->getHeader()->getParent()->hasProfileData()) {
if (auto ProfileTripCount = getLoopEstimatedTripCount(L)) {
if (*ProfileTripCount < FlatLoopTripCountThreshold)
return false;
UP.AllowExpensiveTripCount = true;
UP.Runtime |= PragmaEnableUnroll || PragmaCount > 0 || UserUnrollCount;
if (!UP.Runtime) {
dbgs() << " will not try to unroll loop with runtime trip count "
<< "-unroll-runtime not given\n");
UP.Count = 0;
return false;
if (UP.Count == 0)
UP.Count = UP.DefaultUnrollRuntimeCount;
// Reduce unroll count to be the largest power-of-two factor of
// the original count which satisfies the threshold limit.
while (UP.Count != 0 &&
UCE.getUnrolledLoopSize(UP) > UP.PartialThreshold)
UP.Count >>= 1;
#ifndef NDEBUG
unsigned OrigCount = UP.Count;
if (!UP.AllowRemainder && UP.Count != 0 && (TripMultiple % UP.Count) != 0) {
while (UP.Count != 0 && TripMultiple % UP.Count != 0)
UP.Count >>= 1;
dbgs() << "Remainder loop is restricted (that could architecture "
"specific or because the loop contains a convergent "
"instruction), so unroll count must divide the trip "
"multiple, "
<< TripMultiple << ". Reducing unroll count from " << OrigCount
<< " to " << UP.Count << ".\n");
using namespace ore;
if (unrollCountPragmaValue(L) > 0 && !UP.AllowRemainder)
ORE->emit([&]() {
return OptimizationRemarkMissed(DEBUG_TYPE,
L->getStartLoc(), L->getHeader())
<< "Unable to unroll loop the number of times directed by "
"unroll_count pragma because remainder loop is restricted "
"(that could architecture specific or because the loop "
"contains a convergent instruction) and so must have an "
"unroll "
"count that divides the loop trip multiple of "
<< NV("TripMultiple", TripMultiple) << ". Unrolling instead "
<< NV("UnrollCount", UP.Count) << " time(s).";
if (UP.Count > UP.MaxCount)
UP.Count = UP.MaxCount;
if (MaxTripCount && UP.Count > MaxTripCount)
UP.Count = MaxTripCount;
LLVM_DEBUG(dbgs() << " runtime unrolling with count: " << UP.Count
<< "\n");
if (UP.Count < 2)
UP.Count = 0;
return ExplicitUnroll;
static LoopUnrollResult
tryToUnrollLoop(Loop *L, DominatorTree &DT, LoopInfo *LI, ScalarEvolution &SE,
const TargetTransformInfo &TTI, AssumptionCache &AC,
OptimizationRemarkEmitter &ORE, BlockFrequencyInfo *BFI,
ProfileSummaryInfo *PSI, bool PreserveLCSSA, int OptLevel,
bool OnlyWhenForced, bool ForgetAllSCEV,
std::optional<unsigned> ProvidedCount,
std::optional<unsigned> ProvidedThreshold,
std::optional<bool> ProvidedAllowPartial,
std::optional<bool> ProvidedRuntime,
std::optional<bool> ProvidedUpperBound,
std::optional<bool> ProvidedAllowPeeling,
std::optional<bool> ProvidedAllowProfileBasedPeeling,
std::optional<unsigned> ProvidedFullUnrollMaxCount) {
LLVM_DEBUG(dbgs() << "Loop Unroll: F["
<< L->getHeader()->getParent()->getName() << "] Loop %"
<< L->getHeader()->getName() << "\n");
TransformationMode TM = hasUnrollTransformation(L);
if (TM & TM_Disable)
return LoopUnrollResult::Unmodified;
// If this loop isn't forced to be unrolled, avoid unrolling it when the
// parent loop has an explicit unroll-and-jam pragma. This is to prevent
// automatic unrolling from interfering with the user requested
// transformation.
Loop *ParentL = L->getParentLoop();
if (ParentL != nullptr &&
hasUnrollAndJamTransformation(ParentL) == TM_ForcedByUser &&
hasUnrollTransformation(L) != TM_ForcedByUser) {
LLVM_DEBUG(dbgs() << "Not unrolling loop since parent loop has"
<< " llvm.loop.unroll_and_jam.\n");
return LoopUnrollResult::Unmodified;
// If this loop isn't forced to be unrolled, avoid unrolling it when the
// loop has an explicit unroll-and-jam pragma. This is to prevent automatic
// unrolling from interfering with the user requested transformation.
if (hasUnrollAndJamTransformation(L) == TM_ForcedByUser &&
hasUnrollTransformation(L) != TM_ForcedByUser) {
<< " Not unrolling loop since it has llvm.loop.unroll_and_jam.\n");
return LoopUnrollResult::Unmodified;
if (!L->isLoopSimplifyForm()) {
dbgs() << " Not unrolling loop which is not in loop-simplify form.\n");
return LoopUnrollResult::Unmodified;
// When automatic unrolling is disabled, do not unroll unless overridden for
// this loop.
if (OnlyWhenForced && !(TM & TM_Enable))
return LoopUnrollResult::Unmodified;
bool OptForSize = L->getHeader()->getParent()->hasOptSize();
unsigned NumInlineCandidates;
bool NotDuplicatable;
bool Convergent;
TargetTransformInfo::UnrollingPreferences UP = gatherUnrollingPreferences(
L, SE, TTI, BFI, PSI, ORE, OptLevel, ProvidedThreshold, ProvidedCount,
ProvidedAllowPartial, ProvidedRuntime, ProvidedUpperBound,
TargetTransformInfo::PeelingPreferences PP = gatherPeelingPreferences(
L, SE, TTI, ProvidedAllowPeeling, ProvidedAllowProfileBasedPeeling, true);
// Exit early if unrolling is disabled. For OptForSize, we pick the loop size
// as threshold later on.
if (UP.Threshold == 0 && (!UP.Partial || UP.PartialThreshold == 0) &&
return LoopUnrollResult::Unmodified;
SmallPtrSet<const Value *, 32> EphValues;
CodeMetrics::collectEphemeralValues(L, &AC, EphValues);
InstructionCost LoopSizeIC =
ApproximateLoopSize(L, NumInlineCandidates, NotDuplicatable, Convergent,
TTI, EphValues, UP.BEInsns);
LLVM_DEBUG(dbgs() << " Loop Size = " << LoopSizeIC << "\n");
if (!LoopSizeIC.isValid()) {
LLVM_DEBUG(dbgs() << " Not unrolling loop which contains instructions"
<< " with invalid cost.\n");
return LoopUnrollResult::Unmodified;
unsigned LoopSize = *LoopSizeIC.getValue();
if (NotDuplicatable) {
LLVM_DEBUG(dbgs() << " Not unrolling loop which contains non-duplicatable"
<< " instructions.\n");
return LoopUnrollResult::Unmodified;
// When optimizing for size, use LoopSize + 1 as threshold (we use < Threshold
// later), to (fully) unroll loops, if it does not increase code size.
if (OptForSize)
UP.Threshold = std::max(UP.Threshold, LoopSize + 1);
if (NumInlineCandidates != 0) {
LLVM_DEBUG(dbgs() << " Not unrolling loop with inlinable calls.\n");
return LoopUnrollResult::Unmodified;
// Find the smallest exact trip count for any exit. This is an upper bound
// on the loop trip count, but an exit at an earlier iteration is still
// possible. An unroll by the smallest exact trip count guarantees that all
// branches relating to at least one exit can be eliminated. This is unlike
// the max trip count, which only guarantees that the backedge can be broken.
unsigned TripCount = 0;
unsigned TripMultiple = 1;
SmallVector<BasicBlock *, 8> ExitingBlocks;
for (BasicBlock *ExitingBlock : ExitingBlocks)
if (unsigned TC = SE.getSmallConstantTripCount(L, ExitingBlock))
if (!TripCount || TC < TripCount)
TripCount = TripMultiple = TC;
if (!TripCount) {
// If no exact trip count is known, determine the trip multiple of either
// the loop latch or the single exiting block.
// TODO: Relax for multiple exits.
BasicBlock *ExitingBlock = L->getLoopLatch();
if (!ExitingBlock || !L->isLoopExiting(ExitingBlock))
ExitingBlock = L->getExitingBlock();
if (ExitingBlock)
TripMultiple = SE.getSmallConstantTripMultiple(L, ExitingBlock);
// If the loop contains a convergent operation, the prelude we'd add
// to do the first few instructions before we hit the unrolled loop
// is unsafe -- it adds a control-flow dependency to the convergent
// operation. Therefore restrict remainder loop (try unrolling without).
// TODO: This is quite conservative. In practice, convergent_op()
// is likely to be called unconditionally in the loop. In this
// case, the program would be ill-formed (on most architectures)
// unless n were the same on all threads in a thread group.
// Assuming n is the same on all threads, any kind of unrolling is
// safe. But currently llvm's notion of convergence isn't powerful
// enough to express this.
if (Convergent)
UP.AllowRemainder = false;
// Try to find the trip count upper bound if we cannot find the exact trip
// count.
unsigned MaxTripCount = 0;
bool MaxOrZero = false;
if (!TripCount) {
MaxTripCount = SE.getSmallConstantMaxTripCount(L);
MaxOrZero = SE.isBackedgeTakenCountMaxOrZero(L);
// computeUnrollCount() decides whether it is beneficial to use upper bound to
// fully unroll the loop.
bool UseUpperBound = false;
bool IsCountSetExplicitly = computeUnrollCount(
L, TTI, DT, LI, &AC, SE, EphValues, &ORE, TripCount, MaxTripCount, MaxOrZero,
TripMultiple, LoopSize, UP, PP, UseUpperBound);
if (!UP.Count)
return LoopUnrollResult::Unmodified;
if (PP.PeelCount) {
assert(UP.Count == 1 && "Cannot perform peel and unroll in the same step");
LLVM_DEBUG(dbgs() << "PEELING loop %" << L->getHeader()->getName()
<< " with iteration count " << PP.PeelCount << "!\n");
ORE.emit([&]() {
return OptimizationRemark(DEBUG_TYPE, "Peeled", L->getStartLoc(),
<< " peeled loop by " << ore::NV("PeelCount", PP.PeelCount)
<< " iterations";
ValueToValueMapTy VMap;
if (peelLoop(L, PP.PeelCount, LI, &SE, DT, &AC, PreserveLCSSA, VMap)) {
simplifyLoopAfterUnroll(L, true, LI, &SE, &DT, &AC, &TTI);
// If the loop was peeled, we already "used up" the profile information
// we had, so we don't want to unroll or peel again.
if (PP.PeelProfiledIterations)
return LoopUnrollResult::PartiallyUnrolled;
return LoopUnrollResult::Unmodified;
// At this point, UP.Runtime indicates that run-time unrolling is allowed.
// However, we only want to actually perform it if we don't know the trip
// count and the unroll count doesn't divide the known trip multiple.
// TODO: This decision should probably be pushed up into
// computeUnrollCount().
UP.Runtime &= TripCount == 0 && TripMultiple % UP.Count != 0;
// Save loop properties before it is transformed.
MDNode *OrigLoopID = L->getLoopID();
// Unroll the loop.
Loop *RemainderLoop = nullptr;
LoopUnrollResult UnrollResult = UnrollLoop(
{UP.Count, UP.Force, UP.Runtime, UP.AllowExpensiveTripCount,
UP.UnrollRemainder, ForgetAllSCEV},
LI, &SE, &DT, &AC, &TTI, &ORE, PreserveLCSSA, &RemainderLoop);
if (UnrollResult == LoopUnrollResult::Unmodified)
return LoopUnrollResult::Unmodified;
if (RemainderLoop) {
std::optional<MDNode *> RemainderLoopID =
makeFollowupLoopID(OrigLoopID, {LLVMLoopUnrollFollowupAll,
if (RemainderLoopID)
if (UnrollResult != LoopUnrollResult::FullyUnrolled) {
std::optional<MDNode *> NewLoopID =
makeFollowupLoopID(OrigLoopID, {LLVMLoopUnrollFollowupAll,
if (NewLoopID) {
// Do not setLoopAlreadyUnrolled if loop attributes have been specified
// explicitly.
return UnrollResult;
// If loop has an unroll count pragma or unrolled by explicitly set count
// mark loop as unrolled to prevent unrolling beyond that requested.
if (UnrollResult != LoopUnrollResult::FullyUnrolled && IsCountSetExplicitly)
return UnrollResult;
namespace {
class LoopUnroll : public LoopPass {
static char ID; // Pass ID, replacement for typeid
int OptLevel;
/// If false, use a cost model to determine whether unrolling of a loop is
/// profitable. If true, only loops that explicitly request unrolling via
/// metadata are considered. All other loops are skipped.
bool OnlyWhenForced;
/// If false, when SCEV is invalidated, only forget everything in the
/// top-most loop (call forgetTopMostLoop), of the loop being processed.
/// Otherwise, forgetAllLoops and rebuild when needed next.
bool ForgetAllSCEV;
std::optional<unsigned> ProvidedCount;
std::optional<unsigned> ProvidedThreshold;
std::optional<bool> ProvidedAllowPartial;
std::optional<bool> ProvidedRuntime;
std::optional<bool> ProvidedUpperBound;
std::optional<bool> ProvidedAllowPeeling;
std::optional<bool> ProvidedAllowProfileBasedPeeling;
std::optional<unsigned> ProvidedFullUnrollMaxCount;
LoopUnroll(int OptLevel = 2, bool OnlyWhenForced = false,
bool ForgetAllSCEV = false,
std::optional<unsigned> Threshold = std::nullopt,
std::optional<unsigned> Count = std::nullopt,
std::optional<bool> AllowPartial = std::nullopt,
std::optional<bool> Runtime = std::nullopt,
std::optional<bool> UpperBound = std::nullopt,
std::optional<bool> AllowPeeling = std::nullopt,
std::optional<bool> AllowProfileBasedPeeling = std::nullopt,
std::optional<unsigned> ProvidedFullUnrollMaxCount = std::nullopt)
: LoopPass(ID), OptLevel(OptLevel), OnlyWhenForced(OnlyWhenForced),
ForgetAllSCEV(ForgetAllSCEV), ProvidedCount(std::move(Count)),
ProvidedThreshold(Threshold), ProvidedAllowPartial(AllowPartial),
ProvidedRuntime(Runtime), ProvidedUpperBound(UpperBound),
ProvidedFullUnrollMaxCount(ProvidedFullUnrollMaxCount) {
bool runOnLoop(Loop *L, LPPassManager &LPM) override {
if (skipLoop(L))
return false;
Function &F = *L->getHeader()->getParent();
auto &DT = getAnalysis<DominatorTreeWrapperPass>().getDomTree();
LoopInfo *LI = &getAnalysis<LoopInfoWrapperPass>().getLoopInfo();
ScalarEvolution &SE = getAnalysis<ScalarEvolutionWrapperPass>().getSE();
const TargetTransformInfo &TTI =
auto &AC = getAnalysis<AssumptionCacheTracker>().getAssumptionCache(F);
// For the old PM, we can't use OptimizationRemarkEmitter as an analysis
// pass. Function analyses need to be preserved across loop transformations
// but ORE cannot be preserved (see comment before the pass definition).
OptimizationRemarkEmitter ORE(&F);
bool PreserveLCSSA = mustPreserveAnalysisID(LCSSAID);
LoopUnrollResult Result = tryToUnrollLoop(
L, DT, LI, SE, TTI, AC, ORE, nullptr, nullptr, PreserveLCSSA, OptLevel,
OnlyWhenForced, ForgetAllSCEV, ProvidedCount, ProvidedThreshold,
ProvidedAllowPartial, ProvidedRuntime, ProvidedUpperBound,
ProvidedAllowPeeling, ProvidedAllowProfileBasedPeeling,
if (Result == LoopUnrollResult::FullyUnrolled)
return Result != LoopUnrollResult::Unmodified;
/// This transformation requires natural loop information & requires that
/// loop preheaders be inserted into the CFG...
void getAnalysisUsage(AnalysisUsage &AU) const override {
// FIXME: Loop passes are required to preserve domtree, and for now we just
// recreate dom info if anything gets unrolled.
} // end anonymous namespace
char LoopUnroll::ID = 0;
INITIALIZE_PASS_BEGIN(LoopUnroll, "loop-unroll", "Unroll loops", false, false)
INITIALIZE_PASS_END(LoopUnroll, "loop-unroll", "Unroll loops", false, false)
Pass *llvm::createLoopUnrollPass(int OptLevel, bool OnlyWhenForced,
bool ForgetAllSCEV, int Threshold, int Count,
int AllowPartial, int Runtime, int UpperBound,
int AllowPeeling) {
// TODO: It would make more sense for this function to take the optionals
// directly, but that's dangerous since it would silently break out of tree
// callers.
return new LoopUnroll(
OptLevel, OnlyWhenForced, ForgetAllSCEV,
Threshold == -1 ? std::nullopt : std::optional<unsigned>(Threshold),
Count == -1 ? std::nullopt : std::optional<unsigned>(Count),
AllowPartial == -1 ? std::nullopt : std::optional<bool>(AllowPartial),
Runtime == -1 ? std::nullopt : std::optional<bool>(Runtime),
UpperBound == -1 ? std::nullopt : std::optional<bool>(UpperBound),
AllowPeeling == -1 ? std::nullopt : std::optional<bool>(AllowPeeling));
Pass *llvm::createSimpleLoopUnrollPass(int OptLevel, bool OnlyWhenForced,
bool ForgetAllSCEV) {
return createLoopUnrollPass(OptLevel, OnlyWhenForced, ForgetAllSCEV, -1, -1,
0, 0, 0, 1);
PreservedAnalyses LoopFullUnrollPass::run(Loop &L, LoopAnalysisManager &AM,
LoopStandardAnalysisResults &AR,
LPMUpdater &Updater) {
// For the new PM, we can't use OptimizationRemarkEmitter as an analysis
// pass. Function analyses need to be preserved across loop transformations
// but ORE cannot be preserved (see comment before the pass definition).
OptimizationRemarkEmitter ORE(L.getHeader()->getParent());
// Keep track of the previous loop structure so we can identify new loops
// created by unrolling.
Loop *ParentL = L.getParentLoop();
SmallPtrSet<Loop *, 4> OldLoops;
if (ParentL)
OldLoops.insert(ParentL->begin(), ParentL->end());
OldLoops.insert(AR.LI.begin(), AR.LI.end());
std::string LoopName = std::string(L.getName());
bool Changed =
tryToUnrollLoop(&L, AR.DT, &AR.LI, AR.SE, AR.TTI, AR.AC, ORE,
/*BFI*/ nullptr, /*PSI*/ nullptr,
/*PreserveLCSSA*/ true, OptLevel, OnlyWhenForced,
ForgetSCEV, /*Count*/ std::nullopt,
/*Threshold*/ std::nullopt, /*AllowPartial*/ false,
/*Runtime*/ false, /*UpperBound*/ false,
/*AllowPeeling*/ true,
/*AllowProfileBasedPeeling*/ false,
/*FullUnrollMaxCount*/ std::nullopt) !=
if (!Changed)
return PreservedAnalyses::all();
// The parent must not be damaged by unrolling!
#ifndef NDEBUG
if (ParentL)
// Unrolling can do several things to introduce new loops into a loop nest:
// - Full unrolling clones child loops within the current loop but then
// removes the current loop making all of the children appear to be new
// sibling loops.
// When a new loop appears as a sibling loop after fully unrolling,
// its nesting structure has fundamentally changed and we want to revisit
// it to reflect that.
// When unrolling has removed the current loop, we need to tell the
// infrastructure that it is gone.
// Finally, we support a debugging/testing mode where we revisit child loops
// as well. These are not expected to require further optimizations as either
// they or the loop they were cloned from have been directly visited already.
// But the debugging mode allows us to check this assumption.
bool IsCurrentLoopValid = false;
SmallVector<Loop *, 4> SibLoops;
if (ParentL)
SibLoops.append(ParentL->begin(), ParentL->end());
SibLoops.append(AR.LI.begin(), AR.LI.end());
erase_if(SibLoops, [&](Loop *SibLoop) {
if (SibLoop == &L) {
IsCurrentLoopValid = true;
return true;
// Otherwise erase the loop from the list if it was in the old loops.
return OldLoops.contains(SibLoop);
if (!IsCurrentLoopValid) {
Updater.markLoopAsDeleted(L, LoopName);
} else {
// We can only walk child loops if the current loop remained valid.
if (UnrollRevisitChildLoops) {
// Walk *all* of the child loops.
SmallVector<Loop *, 4> ChildLoops(L.begin(), L.end());
return getLoopPassPreservedAnalyses();
PreservedAnalyses LoopUnrollPass::run(Function &F,
FunctionAnalysisManager &AM) {
auto &LI = AM.getResult<LoopAnalysis>(F);
// There are no loops in the function. Return before computing other expensive
// analyses.
if (LI.empty())
return PreservedAnalyses::all();
auto &SE = AM.getResult<ScalarEvolutionAnalysis>(F);
auto &TTI = AM.getResult<TargetIRAnalysis>(F);
auto &DT = AM.getResult<DominatorTreeAnalysis>(F);
auto &AC = AM.getResult<AssumptionAnalysis>(F);
auto &ORE = AM.getResult<OptimizationRemarkEmitterAnalysis>(F);
LoopAnalysisManager *LAM = nullptr;
if (auto *LAMProxy = AM.getCachedResult<LoopAnalysisManagerFunctionProxy>(F))
LAM = &LAMProxy->getManager();
auto &MAMProxy = AM.getResult<ModuleAnalysisManagerFunctionProxy>(F);
ProfileSummaryInfo *PSI =
auto *BFI = (PSI && PSI->hasProfileSummary()) ?
&AM.getResult<BlockFrequencyAnalysis>(F) : nullptr;
bool Changed = false;
// The unroller requires loops to be in simplified form, and also needs LCSSA.
// Since simplification may add new inner loops, it has to run before the
// legality and profitability checks. This means running the loop unroller
// will simplify all loops, regardless of whether anything end up being
// unrolled.
for (const auto &L : LI) {
Changed |=
simplifyLoop(L, &DT, &LI, &SE, &AC, nullptr, false /* PreserveLCSSA */);
Changed |= formLCSSARecursively(*L, DT, &LI, &SE);
// Add the loop nests in the reverse order of LoopInfo. See method
// declaration.
SmallPriorityWorklist<Loop *, 4> Worklist;
appendLoopsToWorklist(LI, Worklist);
while (!Worklist.empty()) {
// Because the LoopInfo stores the loops in RPO, we walk the worklist
// from back to front so that we work forward across the CFG, which
// for unrolling is only needed to get optimization remarks emitted in
// a forward order.
Loop &L = *Worklist.pop_back_val();
#ifndef NDEBUG
Loop *ParentL = L.getParentLoop();
// Check if the profile summary indicates that the profiled application
// has a huge working set size, in which case we disable peeling to avoid
// bloating it further.
std::optional<bool> LocalAllowPeeling = UnrollOpts.AllowPeeling;
if (PSI && PSI->hasHugeWorkingSetSize())
LocalAllowPeeling = false;
std::string LoopName = std::string(L.getName());
// The API here is quite complex to call and we allow to select some
// flavors of unrolling during construction time (by setting UnrollOpts).
LoopUnrollResult Result = tryToUnrollLoop(
/*PreserveLCSSA*/ true, UnrollOpts.OptLevel, UnrollOpts.OnlyWhenForced,
UnrollOpts.ForgetSCEV, /*Count*/ std::nullopt,
/*Threshold*/ std::nullopt, UnrollOpts.AllowPartial,
UnrollOpts.AllowRuntime, UnrollOpts.AllowUpperBound, LocalAllowPeeling,
UnrollOpts.AllowProfileBasedPeeling, UnrollOpts.FullUnrollMaxCount);
Changed |= Result != LoopUnrollResult::Unmodified;
// The parent must not be damaged by unrolling!
#ifndef NDEBUG
if (Result != LoopUnrollResult::Unmodified && ParentL)
// Clear any cached analysis results for L if we removed it completely.
if (LAM && Result == LoopUnrollResult::FullyUnrolled)
LAM->clear(L, LoopName);
if (!Changed)
return PreservedAnalyses::all();
return getLoopPassPreservedAnalyses();
void LoopUnrollPass::printPipeline(
raw_ostream &OS, function_ref<StringRef(StringRef)> MapClassName2PassName) {
static_cast<PassInfoMixin<LoopUnrollPass> *>(this)->printPipeline(
OS, MapClassName2PassName);
OS << "<";
if (UnrollOpts.AllowPartial != std::nullopt)
OS << (*UnrollOpts.AllowPartial ? "" : "no-") << "partial;";
if (UnrollOpts.AllowPeeling != std::nullopt)
OS << (*UnrollOpts.AllowPeeling ? "" : "no-") << "peeling;";
if (UnrollOpts.AllowRuntime != std::nullopt)
OS << (*UnrollOpts.AllowRuntime ? "" : "no-") << "runtime;";
if (UnrollOpts.AllowUpperBound != std::nullopt)
OS << (*UnrollOpts.AllowUpperBound ? "" : "no-") << "upperbound;";
if (UnrollOpts.AllowProfileBasedPeeling != std::nullopt)
OS << (*UnrollOpts.AllowProfileBasedPeeling ? "" : "no-")
<< "profile-peeling;";
if (UnrollOpts.FullUnrollMaxCount != std::nullopt)
OS << "full-unroll-max=" << UnrollOpts.FullUnrollMaxCount << ";";
OS << "O" << UnrollOpts.OptLevel;
OS << ">";