Viewing file:  MachineScheduler.h (48.31 KB) -rw-r--r--
Select action/file-type:
 (+) |  (+) |  (+) | Code (+) | Session (+) |  (+) | SDB (+) |  (+) |  (+) |  (+) |  (+) |  (+) |
//===- MachineScheduler.h - MachineInstr Scheduling Pass --------*- C++ -*-===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
//
// This file provides an interface for customizing the standard MachineScheduler
// pass. Note that the entire pass may be replaced as follows:
//
// <Target>TargetMachine::createPassConfig(PassManagerBase &PM) {
// PM.substitutePass(&MachineSchedulerID, &CustomSchedulerPassID);
// ...}
//
// The MachineScheduler pass is only responsible for choosing the regions to be
// scheduled. Targets can override the DAG builder and scheduler without
// replacing the pass as follows:
//
// ScheduleDAGInstrs *<Target>PassConfig::
// createMachineScheduler(MachineSchedContext *C) {
// return new CustomMachineScheduler(C);
// }
//
// The default scheduler, ScheduleDAGMILive, builds the DAG and drives list
// scheduling while updating the instruction stream, register pressure, and live
// intervals. Most targets don't need to override the DAG builder and list
// scheduler, but subtargets that require custom scheduling heuristics may
// plugin an alternate MachineSchedStrategy. The strategy is responsible for
// selecting the highest priority node from the list:
//
// ScheduleDAGInstrs *<Target>PassConfig::
// createMachineScheduler(MachineSchedContext *C) {
// return new ScheduleDAGMILive(C, CustomStrategy(C));
// }
//
// The DAG builder can also be customized in a sense by adding DAG mutations
// that will run after DAG building and before list scheduling. DAG mutations
// can adjust dependencies based on target-specific knowledge or add weak edges
// to aid heuristics:
//
// ScheduleDAGInstrs *<Target>PassConfig::
// createMachineScheduler(MachineSchedContext *C) {
// ScheduleDAGMI *DAG = createGenericSchedLive(C);
// DAG->addMutation(new CustomDAGMutation(...));
// return DAG;
// }
//
// A target that supports alternative schedulers can use the
// MachineSchedRegistry to allow command line selection. This can be done by
// implementing the following boilerplate:
//
// static ScheduleDAGInstrs *createCustomMachineSched(MachineSchedContext *C) {
// return new CustomMachineScheduler(C);
// }
// static MachineSchedRegistry
// SchedCustomRegistry("custom", "Run my target's custom scheduler",
// createCustomMachineSched);
//
//
// Finally, subtargets that don't need to implement custom heuristics but would
// like to configure the GenericScheduler's policy for a given scheduler region,
// including scheduling direction and register pressure tracking policy, can do
// this:
//
// void <SubTarget>Subtarget::
// overrideSchedPolicy(MachineSchedPolicy &Policy,
// unsigned NumRegionInstrs) const {
// Policy.<Flag> = true;
// }
//
//===----------------------------------------------------------------------===//
#ifndef LLVM_CODEGEN_MACHINESCHEDULER_H
#define LLVM_CODEGEN_MACHINESCHEDULER_H
#include "llvm/ADT/APInt.h"
#include "llvm/ADT/ArrayRef.h"
#include "llvm/ADT/BitVector.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/ADT/SmallVector.h"
#include "llvm/ADT/StringRef.h"
#include "llvm/ADT/Twine.h"
#include "llvm/CodeGen/MachineBasicBlock.h"
#include "llvm/CodeGen/MachinePassRegistry.h"
#include "llvm/CodeGen/RegisterPressure.h"
#include "llvm/CodeGen/ScheduleDAG.h"
#include "llvm/CodeGen/ScheduleDAGInstrs.h"
#include "llvm/CodeGen/ScheduleDAGMutation.h"
#include "llvm/CodeGen/TargetSchedule.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/ErrorHandling.h"
#include <algorithm>
#include <cassert>
#include <llvm/Support/raw_ostream.h>
#include <memory>
#include <string>
#include <vector>
namespace llvm {
extern cl::opt<bool> ForceTopDown;
extern cl::opt<bool> ForceBottomUp;
extern cl::opt<bool> VerifyScheduling;
#ifndef NDEBUG
extern cl::opt<bool> ViewMISchedDAGs;
extern cl::opt<bool> PrintDAGs;
#else
extern const bool ViewMISchedDAGs;
extern const bool PrintDAGs;
#endif
class AAResults;
class LiveIntervals;
class MachineDominatorTree;
class MachineFunction;
class MachineInstr;
class MachineLoopInfo;
class RegisterClassInfo;
class SchedDFSResult;
class ScheduleHazardRecognizer;
class TargetInstrInfo;
class TargetPassConfig;
class TargetRegisterInfo;
/// MachineSchedContext provides enough context from the MachineScheduler pass
/// for the target to instantiate a scheduler.
struct MachineSchedContext {
MachineFunction *MF = nullptr;
const MachineLoopInfo *MLI = nullptr;
const MachineDominatorTree *MDT = nullptr;
const TargetPassConfig *PassConfig = nullptr;
AAResults *AA = nullptr;
LiveIntervals *LIS = nullptr;
RegisterClassInfo *RegClassInfo;
MachineSchedContext();
MachineSchedContext &operator=(const MachineSchedContext &other) = delete;
MachineSchedContext(const MachineSchedContext &other) = delete;
virtual ~MachineSchedContext();
};
/// MachineSchedRegistry provides a selection of available machine instruction
/// schedulers.
class MachineSchedRegistry
: public MachinePassRegistryNode<
ScheduleDAGInstrs *(*)(MachineSchedContext *)> {
public:
using ScheduleDAGCtor = ScheduleDAGInstrs *(*)(MachineSchedContext *);
// RegisterPassParser requires a (misnamed) FunctionPassCtor type.
using FunctionPassCtor = ScheduleDAGCtor;
static MachinePassRegistry<ScheduleDAGCtor> Registry;
MachineSchedRegistry(const char *N, const char *D, ScheduleDAGCtor C)
: MachinePassRegistryNode(N, D, C) {
Registry.Add(this);
}
~MachineSchedRegistry() { Registry.Remove(this); }
// Accessors.
//
MachineSchedRegistry *getNext() const {
return (MachineSchedRegistry *)MachinePassRegistryNode::getNext();
}
static MachineSchedRegistry *getList() {
return (MachineSchedRegistry *)Registry.getList();
}
static void setListener(MachinePassRegistryListener<FunctionPassCtor> *L) {
Registry.setListener(L);
}
};
class ScheduleDAGMI;
/// Define a generic scheduling policy for targets that don't provide their own
/// MachineSchedStrategy. This can be overriden for each scheduling region
/// before building the DAG.
struct MachineSchedPolicy {
// Allow the scheduler to disable register pressure tracking.
bool ShouldTrackPressure = false;
/// Track LaneMasks to allow reordering of independent subregister writes
/// of the same vreg. \sa MachineSchedStrategy::shouldTrackLaneMasks()
bool ShouldTrackLaneMasks = false;
// Allow the scheduler to force top-down or bottom-up scheduling. If neither
// is true, the scheduler runs in both directions and converges.
bool OnlyTopDown = false;
bool OnlyBottomUp = false;
// Disable heuristic that tries to fetch nodes from long dependency chains
// first.
bool DisableLatencyHeuristic = false;
// Compute DFSResult for use in scheduling heuristics.
bool ComputeDFSResult = false;
MachineSchedPolicy() = default;
};
/// MachineSchedStrategy - Interface to the scheduling algorithm used by
/// ScheduleDAGMI.
///
/// Initialization sequence:
/// initPolicy -> shouldTrackPressure -> initialize(DAG) -> registerRoots
class MachineSchedStrategy {
virtual void anchor();
public:
virtual ~MachineSchedStrategy() = default;
/// Optionally override the per-region scheduling policy.
virtual void initPolicy(MachineBasicBlock::iterator Begin,
MachineBasicBlock::iterator End,
unsigned NumRegionInstrs) {}
virtual void dumpPolicy() const {}
/// Check if pressure tracking is needed before building the DAG and
/// initializing this strategy. Called after initPolicy.
virtual bool shouldTrackPressure() const { return true; }
/// Returns true if lanemasks should be tracked. LaneMask tracking is
/// necessary to reorder independent subregister defs for the same vreg.
/// This has to be enabled in combination with shouldTrackPressure().
virtual bool shouldTrackLaneMasks() const { return false; }
// If this method returns true, handling of the scheduling regions
// themselves (in case of a scheduling boundary in MBB) will be done
// beginning with the topmost region of MBB.
virtual bool doMBBSchedRegionsTopDown() const { return false; }
/// Initialize the strategy after building the DAG for a new region.
virtual void initialize(ScheduleDAGMI *DAG) = 0;
/// Tell the strategy that MBB is about to be processed.
virtual void enterMBB(MachineBasicBlock *MBB) {};
/// Tell the strategy that current MBB is done.
virtual void leaveMBB() {};
/// Notify this strategy that all roots have been released (including those
/// that depend on EntrySU or ExitSU).
virtual void registerRoots() {}
/// Pick the next node to schedule, or return NULL. Set IsTopNode to true to
/// schedule the node at the top of the unscheduled region. Otherwise it will
/// be scheduled at the bottom.
virtual SUnit *pickNode(bool &IsTopNode) = 0;
/// Scheduler callback to notify that a new subtree is scheduled.
virtual void scheduleTree(unsigned SubtreeID) {}
/// Notify MachineSchedStrategy that ScheduleDAGMI has scheduled an
/// instruction and updated scheduled/remaining flags in the DAG nodes.
virtual void schedNode(SUnit *SU, bool IsTopNode) = 0;
/// When all predecessor dependencies have been resolved, free this node for
/// top-down scheduling.
virtual void releaseTopNode(SUnit *SU) = 0;
/// When all successor dependencies have been resolved, free this node for
/// bottom-up scheduling.
virtual void releaseBottomNode(SUnit *SU) = 0;
};
/// ScheduleDAGMI is an implementation of ScheduleDAGInstrs that simply
/// schedules machine instructions according to the given MachineSchedStrategy
/// without much extra book-keeping. This is the common functionality between
/// PreRA and PostRA MachineScheduler.
class ScheduleDAGMI : public ScheduleDAGInstrs {
protected:
AAResults *AA;
LiveIntervals *LIS;
std::unique_ptr<MachineSchedStrategy> SchedImpl;
/// Ordered list of DAG postprocessing steps.
std::vector<std::unique_ptr<ScheduleDAGMutation>> Mutations;
/// The top of the unscheduled zone.
MachineBasicBlock::iterator CurrentTop;
/// The bottom of the unscheduled zone.
MachineBasicBlock::iterator CurrentBottom;
/// Record the next node in a scheduled cluster.
const SUnit *NextClusterPred = nullptr;
const SUnit *NextClusterSucc = nullptr;
#if LLVM_ENABLE_ABI_BREAKING_CHECKS
/// The number of instructions scheduled so far. Used to cut off the
/// scheduler at the point determined by misched-cutoff.
unsigned NumInstrsScheduled = 0;
#endif
public:
ScheduleDAGMI(MachineSchedContext *C, std::unique_ptr<MachineSchedStrategy> S,
bool RemoveKillFlags)
: ScheduleDAGInstrs(*C->MF, C->MLI, RemoveKillFlags), AA(C->AA),
LIS(C->LIS), SchedImpl(std::move(S)) {}
// Provide a vtable anchor
~ScheduleDAGMI() override;
/// If this method returns true, handling of the scheduling regions
/// themselves (in case of a scheduling boundary in MBB) will be done
/// beginning with the topmost region of MBB.
bool doMBBSchedRegionsTopDown() const override {
return SchedImpl->doMBBSchedRegionsTopDown();
}
// Returns LiveIntervals instance for use in DAG mutators and such.
LiveIntervals *getLIS() const { return LIS; }
/// Return true if this DAG supports VReg liveness and RegPressure.
virtual bool hasVRegLiveness() const { return false; }
/// Add a postprocessing step to the DAG builder.
/// Mutations are applied in the order that they are added after normal DAG
/// building and before MachineSchedStrategy initialization.
///
/// ScheduleDAGMI takes ownership of the Mutation object.
void addMutation(std::unique_ptr<ScheduleDAGMutation> Mutation) {
if (Mutation)
Mutations.push_back(std::move(Mutation));
}
MachineBasicBlock::iterator top() const { return CurrentTop; }
MachineBasicBlock::iterator bottom() const { return CurrentBottom; }
/// Implement the ScheduleDAGInstrs interface for handling the next scheduling
/// region. This covers all instructions in a block, while schedule() may only
/// cover a subset.
void enterRegion(MachineBasicBlock *bb,
MachineBasicBlock::iterator begin,
MachineBasicBlock::iterator end,
unsigned regioninstrs) override;
/// Implement ScheduleDAGInstrs interface for scheduling a sequence of
/// reorderable instructions.
void schedule() override;
void startBlock(MachineBasicBlock *bb) override;
void finishBlock() override;
/// Change the position of an instruction within the basic block and update
/// live ranges and region boundary iterators.
void moveInstruction(MachineInstr *MI, MachineBasicBlock::iterator InsertPos);
const SUnit *getNextClusterPred() const { return NextClusterPred; }
const SUnit *getNextClusterSucc() const { return NextClusterSucc; }
void viewGraph(const Twine &Name, const Twine &Title) override;
void viewGraph() override;
protected:
// Top-Level entry points for the schedule() driver...
/// Apply each ScheduleDAGMutation step in order. This allows different
/// instances of ScheduleDAGMI to perform custom DAG postprocessing.
void postProcessDAG();
/// Release ExitSU predecessors and setup scheduler queues.
void initQueues(ArrayRef<SUnit*> TopRoots, ArrayRef<SUnit*> BotRoots);
/// Update scheduler DAG and queues after scheduling an instruction.
void updateQueues(SUnit *SU, bool IsTopNode);
/// Reinsert debug_values recorded in ScheduleDAGInstrs::DbgValues.
void placeDebugValues();
/// dump the scheduled Sequence.
void dumpSchedule() const;
/// Print execution trace of the schedule top-down or bottom-up.
void dumpScheduleTraceTopDown() const;
void dumpScheduleTraceBottomUp() const;
// Lesser helpers...
bool checkSchedLimit();
void findRootsAndBiasEdges(SmallVectorImpl<SUnit*> &TopRoots,
SmallVectorImpl<SUnit*> &BotRoots);
void releaseSucc(SUnit *SU, SDep *SuccEdge);
void releaseSuccessors(SUnit *SU);
void releasePred(SUnit *SU, SDep *PredEdge);
void releasePredecessors(SUnit *SU);
};
/// ScheduleDAGMILive is an implementation of ScheduleDAGInstrs that schedules
/// machine instructions while updating LiveIntervals and tracking regpressure.
class ScheduleDAGMILive : public ScheduleDAGMI {
protected:
RegisterClassInfo *RegClassInfo;
/// Information about DAG subtrees. If DFSResult is NULL, then SchedulerTrees
/// will be empty.
SchedDFSResult *DFSResult = nullptr;
BitVector ScheduledTrees;
MachineBasicBlock::iterator LiveRegionEnd;
/// Maps vregs to the SUnits of their uses in the current scheduling region.
VReg2SUnitMultiMap VRegUses;
// Map each SU to its summary of pressure changes. This array is updated for
// liveness during bottom-up scheduling. Top-down scheduling may proceed but
// has no affect on the pressure diffs.
PressureDiffs SUPressureDiffs;
/// Register pressure in this region computed by initRegPressure.
bool ShouldTrackPressure = false;
bool ShouldTrackLaneMasks = false;
IntervalPressure RegPressure;
RegPressureTracker RPTracker;
/// List of pressure sets that exceed the target's pressure limit before
/// scheduling, listed in increasing set ID order. Each pressure set is paired
/// with its max pressure in the currently scheduled regions.
std::vector<PressureChange> RegionCriticalPSets;
/// The top of the unscheduled zone.
IntervalPressure TopPressure;
RegPressureTracker TopRPTracker;
/// The bottom of the unscheduled zone.
IntervalPressure BotPressure;
RegPressureTracker BotRPTracker;
public:
ScheduleDAGMILive(MachineSchedContext *C,
std::unique_ptr<MachineSchedStrategy> S)
: ScheduleDAGMI(C, std::move(S), /*RemoveKillFlags=*/false),
RegClassInfo(C->RegClassInfo), RPTracker(RegPressure),
TopRPTracker(TopPressure), BotRPTracker(BotPressure) {}
~ScheduleDAGMILive() override;
/// Return true if this DAG supports VReg liveness and RegPressure.
bool hasVRegLiveness() const override { return true; }
/// Return true if register pressure tracking is enabled.
bool isTrackingPressure() const { return ShouldTrackPressure; }
/// Get current register pressure for the top scheduled instructions.
const IntervalPressure &getTopPressure() const { return TopPressure; }
const RegPressureTracker &getTopRPTracker() const { return TopRPTracker; }
/// Get current register pressure for the bottom scheduled instructions.
const IntervalPressure &getBotPressure() const { return BotPressure; }
const RegPressureTracker &getBotRPTracker() const { return BotRPTracker; }
/// Get register pressure for the entire scheduling region before scheduling.
const IntervalPressure &getRegPressure() const { return RegPressure; }
const std::vector<PressureChange> &getRegionCriticalPSets() const {
return RegionCriticalPSets;
}
PressureDiff &getPressureDiff(const SUnit *SU) {
return SUPressureDiffs[SU->NodeNum];
}
const PressureDiff &getPressureDiff(const SUnit *SU) const {
return SUPressureDiffs[SU->NodeNum];
}
/// Compute a DFSResult after DAG building is complete, and before any
/// queue comparisons.
void computeDFSResult();
/// Return a non-null DFS result if the scheduling strategy initialized it.
const SchedDFSResult *getDFSResult() const { return DFSResult; }
BitVector &getScheduledTrees() { return ScheduledTrees; }
/// Implement the ScheduleDAGInstrs interface for handling the next scheduling
/// region. This covers all instructions in a block, while schedule() may only
/// cover a subset.
void enterRegion(MachineBasicBlock *bb,
MachineBasicBlock::iterator begin,
MachineBasicBlock::iterator end,
unsigned regioninstrs) override;
/// Implement ScheduleDAGInstrs interface for scheduling a sequence of
/// reorderable instructions.
void schedule() override;
/// Compute the cyclic critical path through the DAG.
unsigned computeCyclicCriticalPath();
void dump() const override;
protected:
// Top-Level entry points for the schedule() driver...
/// Call ScheduleDAGInstrs::buildSchedGraph with register pressure tracking
/// enabled. This sets up three trackers. RPTracker will cover the entire DAG
/// region, TopTracker and BottomTracker will be initialized to the top and
/// bottom of the DAG region without covereing any unscheduled instruction.
void buildDAGWithRegPressure();
/// Release ExitSU predecessors and setup scheduler queues. Re-position
/// the Top RP tracker in case the region beginning has changed.
void initQueues(ArrayRef<SUnit*> TopRoots, ArrayRef<SUnit*> BotRoots);
/// Move an instruction and update register pressure.
void scheduleMI(SUnit *SU, bool IsTopNode);
// Lesser helpers...
void initRegPressure();
void updatePressureDiffs(ArrayRef<RegisterMaskPair> LiveUses);
void updateScheduledPressure(const SUnit *SU,
const std::vector<unsigned> &NewMaxPressure);
void collectVRegUses(SUnit &SU);
};
//===----------------------------------------------------------------------===//
///
/// Helpers for implementing custom MachineSchedStrategy classes. These take
/// care of the book-keeping associated with list scheduling heuristics.
///
//===----------------------------------------------------------------------===//
/// ReadyQueue encapsulates vector of "ready" SUnits with basic convenience
/// methods for pushing and removing nodes. ReadyQueue's are uniquely identified
/// by an ID. SUnit::NodeQueueId is a mask of the ReadyQueues the SUnit is in.
///
/// This is a convenience class that may be used by implementations of
/// MachineSchedStrategy.
class ReadyQueue {
unsigned ID;
std::string Name;
std::vector<SUnit*> Queue;
public:
ReadyQueue(unsigned id, const Twine &name): ID(id), Name(name.str()) {}
unsigned getID() const { return ID; }
StringRef getName() const { return Name; }
// SU is in this queue if it's NodeQueueID is a superset of this ID.
bool isInQueue(SUnit *SU) const { return (SU->NodeQueueId & ID); }
bool empty() const { return Queue.empty(); }
void clear() { Queue.clear(); }
unsigned size() const { return Queue.size(); }
using iterator = std::vector<SUnit*>::iterator;
iterator begin() { return Queue.begin(); }
iterator end() { return Queue.end(); }
ArrayRef<SUnit*> elements() { return Queue; }
iterator find(SUnit *SU) { return llvm::find(Queue, SU); }
void push(SUnit *SU) {
Queue.push_back(SU);
SU->NodeQueueId |= ID;
}
iterator remove(iterator I) {
(*I)->NodeQueueId &= ~ID;
*I = Queue.back();
unsigned idx = I - Queue.begin();
Queue.pop_back();
return Queue.begin() + idx;
}
void dump() const;
};
/// Summarize the unscheduled region.
struct SchedRemainder {
// Critical path through the DAG in expected latency.
unsigned CriticalPath;
unsigned CyclicCritPath;
// Scaled count of micro-ops left to schedule.
unsigned RemIssueCount;
bool IsAcyclicLatencyLimited;
// Unscheduled resources
SmallVector<unsigned, 16> RemainingCounts;
SchedRemainder() { reset(); }
void reset() {
CriticalPath = 0;
CyclicCritPath = 0;
RemIssueCount = 0;
IsAcyclicLatencyLimited = false;
RemainingCounts.clear();
}
void init(ScheduleDAGMI *DAG, const TargetSchedModel *SchedModel);
};
/// ResourceSegments are a collection of intervals closed on the
/// left and opened on the right:
///
/// list{ [a1, b1), [a2, b2), ..., [a_N, b_N) }
///
/// The collection has the following properties:
///
/// 1. The list is ordered: a_i < b_i and b_i < a_(i+1)
///
/// 2. The intervals in the collection do not intersect each other.
///
/// A \ref ResourceSegments instance represents the cycle
/// reservation history of the instance of and individual resource.
class ResourceSegments {
public:
/// Represents an interval of discrete integer values closed on
/// the left and open on the right: [a, b).
typedef std::pair<int64_t, int64_t> IntervalTy;
/// Adds an interval [a, b) to the collection of the instance.
///
/// When adding [a, b[ to the collection, the operation merges the
/// adjacent intervals. For example
///
/// 0 1 2 3 4 5 6 7 8 9 10
/// [-----) [--) [--)
/// + [--)
/// = [-----------) [--)
///
/// To be able to debug duplicate resource usage, the function has
/// assertion that checks that no interval should be added if it
/// overlaps any of the intervals in the collection. We can
/// require this because by definition a \ref ResourceSegments is
/// attached only to an individual resource instance.
void add(IntervalTy A, const unsigned CutOff = 10);
public:
/// Checks whether intervals intersect.
static bool intersects(IntervalTy A, IntervalTy B);
/// These function return the interval used by a resource in bottom and top
/// scheduling.
///
/// Consider an instruction that uses resources X0, X1 and X2 as follows:
///
/// X0 X1 X1 X2 +--------+-------------+--------------+
/// |Resource|AcquireAtCycle|ReleaseAtCycle|
/// +--------+-------------+--------------+
/// | X0 | 0 | 1 |
/// +--------+-------------+--------------+
/// | X1 | 1 | 3 |
/// +--------+-------------+--------------+
/// | X2 | 3 | 4 |
/// +--------+-------------+--------------+
///
/// If we can schedule the instruction at cycle C, we need to
/// compute the interval of the resource as follows:
///
/// # TOP DOWN SCHEDULING
///
/// Cycles scheduling flows to the _right_, in the same direction
/// of time.
///
/// C 1 2 3 4 5 ...
/// ------|------|------|------|------|------|----->
/// X0 X1 X1 X2 ---> direction of time
/// X0 [C, C+1)
/// X1 [C+1, C+3)
/// X2 [C+3, C+4)
///
/// Therefore, the formula to compute the interval for a resource
/// of an instruction that can be scheduled at cycle C in top-down
/// scheduling is:
///
/// [C+AcquireAtCycle, C+ReleaseAtCycle)
///
///
/// # BOTTOM UP SCHEDULING
///
/// Cycles scheduling flows to the _left_, in opposite direction
/// of time.
///
/// In bottom up scheduling, the scheduling happens in opposite
/// direction to the execution of the cycles of the
/// instruction. When the instruction is scheduled at cycle `C`,
/// the resources are allocated in the past relative to `C`:
///
/// 2 1 C -1 -2 -3 -4 -5 ...
/// <-----|------|------|------|------|------|------|------|---
/// X0 X1 X1 X2 ---> direction of time
/// X0 (C+1, C]
/// X1 (C, C-2]
/// X2 (C-2, C-3]
///
/// Therefore, the formula to compute the interval for a resource
/// of an instruction that can be scheduled at cycle C in bottom-up
/// scheduling is:
///
/// [C-ReleaseAtCycle+1, C-AcquireAtCycle+1)
///
///
/// NOTE: In both cases, the number of cycles booked by a
/// resources is the value (ReleaseAtCycle - AcquireAtCycle).
static IntervalTy getResourceIntervalBottom(unsigned C, unsigned AcquireAtCycle,
unsigned ReleaseAtCycle) {
return std::make_pair<long, long>((long)C - (long)ReleaseAtCycle + 1L,
(long)C - (long)AcquireAtCycle + 1L);
}
static IntervalTy getResourceIntervalTop(unsigned C, unsigned AcquireAtCycle,
unsigned ReleaseAtCycle) {
return std::make_pair<long, long>((long)C + (long)AcquireAtCycle,
(long)C + (long)ReleaseAtCycle);
}
private:
/// Finds the first cycle in which a resource can be allocated.
///
/// The function uses the \param IntervalBuider [*] to build a
/// resource interval [a, b[ out of the input parameters \param
/// CurrCycle, \param AcquireAtCycle and \param ReleaseAtCycle.
///
/// The function then loops through the intervals in the ResourceSegments
/// and shifts the interval [a, b[ and the ReturnCycle to the
/// right until there is no intersection between the intervals of
/// the \ref ResourceSegments instance and the new shifted [a, b[. When
/// this condition is met, the ReturnCycle (which
/// correspond to the cycle in which the resource can be
/// allocated) is returned.
///
/// c = CurrCycle in input
/// c 1 2 3 4 5 6 7 8 9 10 ... ---> (time
/// flow)
/// ResourceSegments... [---) [-------) [-----------)
/// c [1 3[ -> AcquireAtCycle=1, ReleaseAtCycle=3
/// ++c [1 3)
/// ++c [1 3)
/// ++c [1 3)
/// ++c [1 3)
/// ++c [1 3) ---> returns c
/// incremented by 5 (c+5)
///
///
/// Notice that for bottom-up scheduling the diagram is slightly
/// different because the current cycle c is always on the right
/// of the interval [a, b) (see \ref
/// `getResourceIntervalBottom`). This is because the cycle
/// increments for bottom-up scheduling moved in the direction
/// opposite to the direction of time:
///
/// --------> direction of time.
/// XXYZZZ (resource usage)
/// --------> direction of top-down execution cycles.
/// <-------- direction of bottom-up execution cycles.
///
/// Even though bottom-up scheduling moves against the flow of
/// time, the algorithm used to find the first free slot in between
/// intervals is the same as for top-down scheduling.
///
/// [*] See \ref `getResourceIntervalTop` and
/// \ref `getResourceIntervalBottom` to see how such resource intervals
/// are built.
unsigned getFirstAvailableAt(
unsigned CurrCycle, unsigned AcquireAtCycle, unsigned ReleaseAtCycle,
std::function<IntervalTy(unsigned, unsigned, unsigned)> IntervalBuilder)
const;
public:
/// getFirstAvailableAtFromBottom and getFirstAvailableAtFromTop
/// should be merged in a single function in which a function that
/// creates the `NewInterval` is passed as a parameter.
unsigned getFirstAvailableAtFromBottom(unsigned CurrCycle,
unsigned AcquireAtCycle,
unsigned ReleaseAtCycle) const {
return getFirstAvailableAt(CurrCycle, AcquireAtCycle, ReleaseAtCycle,
getResourceIntervalBottom);
}
unsigned getFirstAvailableAtFromTop(unsigned CurrCycle,
unsigned AcquireAtCycle,
unsigned ReleaseAtCycle) const {
return getFirstAvailableAt(CurrCycle, AcquireAtCycle, ReleaseAtCycle,
getResourceIntervalTop);
}
private:
std::list<IntervalTy> _Intervals;
/// Merge all adjacent intervals in the collection. For all pairs
/// of adjacient intervals, it performs [a, b) + [b, c) -> [a, c).
///
/// Before performing the merge operation, the intervals are
/// sorted with \ref sort_predicate.
void sortAndMerge();
public:
// constructor for empty set
explicit ResourceSegments(){};
bool empty() const { return _Intervals.empty(); }
explicit ResourceSegments(const std::list<IntervalTy> &Intervals)
: _Intervals(Intervals) {
sortAndMerge();
}
friend bool operator==(const ResourceSegments &c1,
const ResourceSegments &c2) {
return c1._Intervals == c2._Intervals;
}
friend llvm::raw_ostream &operator<<(llvm::raw_ostream &os,
const ResourceSegments &Segments) {
os << "{ ";
for (auto p : Segments._Intervals)
os << "[" << p.first << ", " << p.second << "), ";
os << "}\n";
return os;
}
};
/// Each Scheduling boundary is associated with ready queues. It tracks the
/// current cycle in the direction of movement, and maintains the state
/// of "hazards" and other interlocks at the current cycle.
class SchedBoundary {
public:
/// SUnit::NodeQueueId: 0 (none), 1 (top), 2 (bot), 3 (both)
enum {
TopQID = 1,
BotQID = 2,
LogMaxQID = 2
};
ScheduleDAGMI *DAG = nullptr;
const TargetSchedModel *SchedModel = nullptr;
SchedRemainder *Rem = nullptr;
ReadyQueue Available;
ReadyQueue Pending;
ScheduleHazardRecognizer *HazardRec = nullptr;
private:
/// True if the pending Q should be checked/updated before scheduling another
/// instruction.
bool CheckPending;
/// Number of cycles it takes to issue the instructions scheduled in this
/// zone. It is defined as: scheduled-micro-ops / issue-width + stalls.
/// See getStalls().
unsigned CurrCycle;
/// Micro-ops issued in the current cycle
unsigned CurrMOps;
/// MinReadyCycle - Cycle of the soonest available instruction.
unsigned MinReadyCycle;
// The expected latency of the critical path in this scheduled zone.
unsigned ExpectedLatency;
// The latency of dependence chains leading into this zone.
// For each node scheduled bottom-up: DLat = max DLat, N.Depth.
// For each cycle scheduled: DLat -= 1.
unsigned DependentLatency;
/// Count the scheduled (issued) micro-ops that can be retired by
/// time=CurrCycle assuming the first scheduled instr is retired at time=0.
unsigned RetiredMOps;
// Count scheduled resources that have been executed. Resources are
// considered executed if they become ready in the time that it takes to
// saturate any resource including the one in question. Counts are scaled
// for direct comparison with other resources. Counts can be compared with
// MOps * getMicroOpFactor and Latency * getLatencyFactor.
SmallVector<unsigned, 16> ExecutedResCounts;
/// Cache the max count for a single resource.
unsigned MaxExecutedResCount;
// Cache the critical resources ID in this scheduled zone.
unsigned ZoneCritResIdx;
// Is the scheduled region resource limited vs. latency limited.
bool IsResourceLimited;
public:
private:
/// Record how resources have been allocated across the cycles of
/// the execution.
std::map<unsigned, ResourceSegments> ReservedResourceSegments;
std::vector<unsigned> ReservedCycles;
/// For each PIdx, stores first index into ReservedResourceSegments that
/// corresponds to it.
///
/// For example, consider the following 3 resources (ResourceCount =
/// 3):
///
/// +------------+--------+
/// |ResourceName|NumUnits|
/// +------------+--------+
/// | X | 2 |
/// +------------+--------+
/// | Y | 3 |
/// +------------+--------+
/// | Z | 1 |
/// +------------+--------+
///
/// In this case, the total number of resource instances is 6. The
/// vector \ref ReservedResourceSegments will have a slot for each instance.
/// The vector \ref ReservedCyclesIndex will track at what index the first
/// instance of the resource is found in the vector of \ref
/// ReservedResourceSegments:
///
/// Indexes of instances in
/// ReservedResourceSegments
///
/// 0 1 2 3 4 5
/// ReservedCyclesIndex[0] = 0; [X0, X1,
/// ReservedCyclesIndex[1] = 2; Y0, Y1, Y2
/// ReservedCyclesIndex[2] = 5; Z
SmallVector<unsigned, 16> ReservedCyclesIndex;
// For each PIdx, stores the resource group IDs of its subunits
SmallVector<APInt, 16> ResourceGroupSubUnitMasks;
#if LLVM_ENABLE_ABI_BREAKING_CHECKS
// Remember the greatest possible stall as an upper bound on the number of
// times we should retry the pending queue because of a hazard.
unsigned MaxObservedStall;
#endif
public:
/// Pending queues extend the ready queues with the same ID and the
/// PendingFlag set.
SchedBoundary(unsigned ID, const Twine &Name):
Available(ID, Name+".A"), Pending(ID << LogMaxQID, Name+".P") {
reset();
}
SchedBoundary &operator=(const SchedBoundary &other) = delete;
SchedBoundary(const SchedBoundary &other) = delete;
~SchedBoundary();
void reset();
void init(ScheduleDAGMI *dag, const TargetSchedModel *smodel,
SchedRemainder *rem);
bool isTop() const {
return Available.getID() == TopQID;
}
/// Number of cycles to issue the instructions scheduled in this zone.
unsigned getCurrCycle() const { return CurrCycle; }
/// Micro-ops issued in the current cycle
unsigned getCurrMOps() const { return CurrMOps; }
// The latency of dependence chains leading into this zone.
unsigned getDependentLatency() const { return DependentLatency; }
/// Get the number of latency cycles "covered" by the scheduled
/// instructions. This is the larger of the critical path within the zone
/// and the number of cycles required to issue the instructions.
unsigned getScheduledLatency() const {
return std::max(ExpectedLatency, CurrCycle);
}
unsigned getUnscheduledLatency(SUnit *SU) const {
return isTop() ? SU->getHeight() : SU->getDepth();
}
unsigned getResourceCount(unsigned ResIdx) const {
return ExecutedResCounts[ResIdx];
}
/// Get the scaled count of scheduled micro-ops and resources, including
/// executed resources.
unsigned getCriticalCount() const {
if (!ZoneCritResIdx)
return RetiredMOps * SchedModel->getMicroOpFactor();
return getResourceCount(ZoneCritResIdx);
}
/// Get a scaled count for the minimum execution time of the scheduled
/// micro-ops that are ready to execute by getExecutedCount. Notice the
/// feedback loop.
unsigned getExecutedCount() const {
return std::max(CurrCycle * SchedModel->getLatencyFactor(),
MaxExecutedResCount);
}
unsigned getZoneCritResIdx() const { return ZoneCritResIdx; }
// Is the scheduled region resource limited vs. latency limited.
bool isResourceLimited() const { return IsResourceLimited; }
/// Get the difference between the given SUnit's ready time and the current
/// cycle.
unsigned getLatencyStallCycles(SUnit *SU);
unsigned getNextResourceCycleByInstance(unsigned InstanceIndex,
unsigned ReleaseAtCycle,
unsigned AcquireAtCycle);
std::pair<unsigned, unsigned> getNextResourceCycle(const MCSchedClassDesc *SC,
unsigned PIdx,
unsigned ReleaseAtCycle,
unsigned AcquireAtCycle);
bool isUnbufferedGroup(unsigned PIdx) const {
return SchedModel->getProcResource(PIdx)->SubUnitsIdxBegin &&
!SchedModel->getProcResource(PIdx)->BufferSize;
}
bool checkHazard(SUnit *SU);
unsigned findMaxLatency(ArrayRef<SUnit*> ReadySUs);
unsigned getOtherResourceCount(unsigned &OtherCritIdx);
/// Release SU to make it ready. If it's not in hazard, remove it from
/// pending queue (if already in) and push into available queue.
/// Otherwise, push the SU into pending queue.
///
/// @param SU The unit to be released.
/// @param ReadyCycle Until which cycle the unit is ready.
/// @param InPQueue Whether SU is already in pending queue.
/// @param Idx Position offset in pending queue (if in it).
void releaseNode(SUnit *SU, unsigned ReadyCycle, bool InPQueue,
unsigned Idx = 0);
void bumpCycle(unsigned NextCycle);
void incExecutedResources(unsigned PIdx, unsigned Count);
unsigned countResource(const MCSchedClassDesc *SC, unsigned PIdx,
unsigned Cycles, unsigned ReadyCycle,
unsigned StartAtCycle);
void bumpNode(SUnit *SU);
void releasePending();
void removeReady(SUnit *SU);
/// Call this before applying any other heuristics to the Available queue.
/// Updates the Available/Pending Q's if necessary and returns the single
/// available instruction, or NULL if there are multiple candidates.
SUnit *pickOnlyChoice();
/// Dump the state of the information that tracks resource usage.
void dumpReservedCycles() const;
void dumpScheduledState() const;
};
/// Base class for GenericScheduler. This class maintains information about
/// scheduling candidates based on TargetSchedModel making it easy to implement
/// heuristics for either preRA or postRA scheduling.
class GenericSchedulerBase : public MachineSchedStrategy {
public:
/// Represent the type of SchedCandidate found within a single queue.
/// pickNodeBidirectional depends on these listed by decreasing priority.
enum CandReason : uint8_t {
NoCand, Only1, PhysReg, RegExcess, RegCritical, Stall, Cluster, Weak,
RegMax, ResourceReduce, ResourceDemand, BotHeightReduce, BotPathReduce,
TopDepthReduce, TopPathReduce, NextDefUse, NodeOrder};
#ifndef NDEBUG
static const char *getReasonStr(GenericSchedulerBase::CandReason Reason);
#endif
/// Policy for scheduling the next instruction in the candidate's zone.
struct CandPolicy {
bool ReduceLatency = false;
unsigned ReduceResIdx = 0;
unsigned DemandResIdx = 0;
CandPolicy() = default;
bool operator==(const CandPolicy &RHS) const {
return ReduceLatency == RHS.ReduceLatency &&
ReduceResIdx == RHS.ReduceResIdx &&
DemandResIdx == RHS.DemandResIdx;
}
bool operator!=(const CandPolicy &RHS) const {
return !(*this == RHS);
}
};
/// Status of an instruction's critical resource consumption.
struct SchedResourceDelta {
// Count critical resources in the scheduled region required by SU.
unsigned CritResources = 0;
// Count critical resources from another region consumed by SU.
unsigned DemandedResources = 0;
SchedResourceDelta() = default;
bool operator==(const SchedResourceDelta &RHS) const {
return CritResources == RHS.CritResources
&& DemandedResources == RHS.DemandedResources;
}
bool operator!=(const SchedResourceDelta &RHS) const {
return !operator==(RHS);
}
};
/// Store the state used by GenericScheduler heuristics, required for the
/// lifetime of one invocation of pickNode().
struct SchedCandidate {
CandPolicy Policy;
// The best SUnit candidate.
SUnit *SU;
// The reason for this candidate.
CandReason Reason;
// Whether this candidate should be scheduled at top/bottom.
bool AtTop;
// Register pressure values for the best candidate.
RegPressureDelta RPDelta;
// Critical resource consumption of the best candidate.
SchedResourceDelta ResDelta;
SchedCandidate() { reset(CandPolicy()); }
SchedCandidate(const CandPolicy &Policy) { reset(Policy); }
void reset(const CandPolicy &NewPolicy) {
Policy = NewPolicy;
SU = nullptr;
Reason = NoCand;
AtTop = false;
RPDelta = RegPressureDelta();
ResDelta = SchedResourceDelta();
}
bool isValid() const { return SU; }
// Copy the status of another candidate without changing policy.
void setBest(SchedCandidate &Best) {
assert(Best.Reason != NoCand && "uninitialized Sched candidate");
SU = Best.SU;
Reason = Best.Reason;
AtTop = Best.AtTop;
RPDelta = Best.RPDelta;
ResDelta = Best.ResDelta;
}
void initResourceDelta(const ScheduleDAGMI *DAG,
const TargetSchedModel *SchedModel);
};
protected:
const MachineSchedContext *Context;
const TargetSchedModel *SchedModel = nullptr;
const TargetRegisterInfo *TRI = nullptr;
SchedRemainder Rem;
GenericSchedulerBase(const MachineSchedContext *C) : Context(C) {}
void setPolicy(CandPolicy &Policy, bool IsPostRA, SchedBoundary &CurrZone,
SchedBoundary *OtherZone);
#ifndef NDEBUG
void traceCandidate(const SchedCandidate &Cand);
#endif
private:
bool shouldReduceLatency(const CandPolicy &Policy, SchedBoundary &CurrZone,
bool ComputeRemLatency, unsigned &RemLatency) const;
};
// Utility functions used by heuristics in tryCandidate().
bool tryLess(int TryVal, int CandVal,
GenericSchedulerBase::SchedCandidate &TryCand,
GenericSchedulerBase::SchedCandidate &Cand,
GenericSchedulerBase::CandReason Reason);
bool tryGreater(int TryVal, int CandVal,
GenericSchedulerBase::SchedCandidate &TryCand,
GenericSchedulerBase::SchedCandidate &Cand,
GenericSchedulerBase::CandReason Reason);
bool tryLatency(GenericSchedulerBase::SchedCandidate &TryCand,
GenericSchedulerBase::SchedCandidate &Cand,
SchedBoundary &Zone);
bool tryPressure(const PressureChange &TryP,
const PressureChange &CandP,
GenericSchedulerBase::SchedCandidate &TryCand,
GenericSchedulerBase::SchedCandidate &Cand,
GenericSchedulerBase::CandReason Reason,
const TargetRegisterInfo *TRI,
const MachineFunction &MF);
unsigned getWeakLeft(const SUnit *SU, bool isTop);
int biasPhysReg(const SUnit *SU, bool isTop);
/// GenericScheduler shrinks the unscheduled zone using heuristics to balance
/// the schedule.
class GenericScheduler : public GenericSchedulerBase {
public:
GenericScheduler(const MachineSchedContext *C):
GenericSchedulerBase(C), Top(SchedBoundary::TopQID, "TopQ"),
Bot(SchedBoundary::BotQID, "BotQ") {}
void initPolicy(MachineBasicBlock::iterator Begin,
MachineBasicBlock::iterator End,
unsigned NumRegionInstrs) override;
void dumpPolicy() const override;
bool shouldTrackPressure() const override {
return RegionPolicy.ShouldTrackPressure;
}
bool shouldTrackLaneMasks() const override {
return RegionPolicy.ShouldTrackLaneMasks;
}
void initialize(ScheduleDAGMI *dag) override;
SUnit *pickNode(bool &IsTopNode) override;
void schedNode(SUnit *SU, bool IsTopNode) override;
void releaseTopNode(SUnit *SU) override {
if (SU->isScheduled)
return;
Top.releaseNode(SU, SU->TopReadyCycle, false);
TopCand.SU = nullptr;
}
void releaseBottomNode(SUnit *SU) override {
if (SU->isScheduled)
return;
Bot.releaseNode(SU, SU->BotReadyCycle, false);
BotCand.SU = nullptr;
}
void registerRoots() override;
protected:
ScheduleDAGMILive *DAG = nullptr;
MachineSchedPolicy RegionPolicy;
// State of the top and bottom scheduled instruction boundaries.
SchedBoundary Top;
SchedBoundary Bot;
/// Candidate last picked from Top boundary.
SchedCandidate TopCand;
/// Candidate last picked from Bot boundary.
SchedCandidate BotCand;
void checkAcyclicLatency();
void initCandidate(SchedCandidate &Cand, SUnit *SU, bool AtTop,
const RegPressureTracker &RPTracker,
RegPressureTracker &TempTracker);
virtual bool tryCandidate(SchedCandidate &Cand, SchedCandidate &TryCand,
SchedBoundary *Zone) const;
SUnit *pickNodeBidirectional(bool &IsTopNode);
void pickNodeFromQueue(SchedBoundary &Zone,
const CandPolicy &ZonePolicy,
const RegPressureTracker &RPTracker,
SchedCandidate &Candidate);
void reschedulePhysReg(SUnit *SU, bool isTop);
};
/// PostGenericScheduler - Interface to the scheduling algorithm used by
/// ScheduleDAGMI.
///
/// Callbacks from ScheduleDAGMI:
/// initPolicy -> initialize(DAG) -> registerRoots -> pickNode ...
class PostGenericScheduler : public GenericSchedulerBase {
protected:
ScheduleDAGMI *DAG = nullptr;
SchedBoundary Top;
SchedBoundary Bot;
MachineSchedPolicy RegionPolicy;
/// Candidate last picked from Top boundary.
SchedCandidate TopCand;
/// Candidate last picked from Bot boundary.
SchedCandidate BotCand;
public:
PostGenericScheduler(const MachineSchedContext *C)
: GenericSchedulerBase(C), Top(SchedBoundary::TopQID, "TopQ"),
Bot(SchedBoundary::BotQID, "BotQ") {}
~PostGenericScheduler() override = default;
void initPolicy(MachineBasicBlock::iterator Begin,
MachineBasicBlock::iterator End,
unsigned NumRegionInstrs) override;
/// PostRA scheduling does not track pressure.
bool shouldTrackPressure() const override { return false; }
void initialize(ScheduleDAGMI *Dag) override;
void registerRoots() override;
SUnit *pickNode(bool &IsTopNode) override;
SUnit *pickNodeBidirectional(bool &IsTopNode);
void scheduleTree(unsigned SubtreeID) override {
llvm_unreachable("PostRA scheduler does not support subtree analysis.");
}
void schedNode(SUnit *SU, bool IsTopNode) override;
void releaseTopNode(SUnit *SU) override {
if (SU->isScheduled)
return;
Top.releaseNode(SU, SU->TopReadyCycle, false);
TopCand.SU = nullptr;
}
void releaseBottomNode(SUnit *SU) override {
if (SU->isScheduled)
return;
Bot.releaseNode(SU, SU->BotReadyCycle, false);
BotCand.SU = nullptr;
}
protected:
virtual bool tryCandidate(SchedCandidate &Cand, SchedCandidate &TryCand);
void pickNodeFromQueue(SchedBoundary &Zone, SchedCandidate &Cand);
};
/// Create the standard converging machine scheduler. This will be used as the
/// default scheduler if the target does not set a default.
/// Adds default DAG mutations.
ScheduleDAGMILive *createGenericSchedLive(MachineSchedContext *C);
/// Create a generic scheduler with no vreg liveness or DAG mutation passes.
ScheduleDAGMI *createGenericSchedPostRA(MachineSchedContext *C);
/// If ReorderWhileClustering is set to true, no attempt will be made to
/// reduce reordering due to store clustering.
std::unique_ptr<ScheduleDAGMutation>
createLoadClusterDAGMutation(const TargetInstrInfo *TII,
const TargetRegisterInfo *TRI,
bool ReorderWhileClustering = false);
/// If ReorderWhileClustering is set to true, no attempt will be made to
/// reduce reordering due to store clustering.
std::unique_ptr<ScheduleDAGMutation>
createStoreClusterDAGMutation(const TargetInstrInfo *TII,
const TargetRegisterInfo *TRI,
bool ReorderWhileClustering = false);
std::unique_ptr<ScheduleDAGMutation>
createCopyConstrainDAGMutation(const TargetInstrInfo *TII,
const TargetRegisterInfo *TRI);
} // end namespace llvm
#endif // LLVM_CODEGEN_MACHINESCHEDULER_H
|