1 //===- MachineScheduler.cpp - Machine Instruction Scheduler ---------------===//
3 // The LLVM Compiler Infrastructure
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
8 //===----------------------------------------------------------------------===//
10 // MachineScheduler schedules machine instructions after phi elimination. It
11 // preserves LiveIntervals so it can be invoked before register allocation.
13 //===----------------------------------------------------------------------===//
15 #define DEBUG_TYPE "misched"
17 #include "llvm/CodeGen/MachineScheduler.h"
18 #include "llvm/ADT/OwningPtr.h"
19 #include "llvm/ADT/PriorityQueue.h"
20 #include "llvm/Analysis/AliasAnalysis.h"
21 #include "llvm/CodeGen/LiveIntervalAnalysis.h"
22 #include "llvm/CodeGen/MachineDominators.h"
23 #include "llvm/CodeGen/MachineLoopInfo.h"
24 #include "llvm/CodeGen/Passes.h"
25 #include "llvm/CodeGen/RegisterClassInfo.h"
26 #include "llvm/CodeGen/ScheduleDFS.h"
27 #include "llvm/CodeGen/ScheduleHazardRecognizer.h"
28 #include "llvm/Support/CommandLine.h"
29 #include "llvm/Support/Debug.h"
30 #include "llvm/Support/ErrorHandling.h"
31 #include "llvm/Support/GraphWriter.h"
32 #include "llvm/Support/raw_ostream.h"
33 #include "llvm/Target/TargetInstrInfo.h"
39 cl::opt<bool> ForceTopDown("misched-topdown", cl::Hidden,
40 cl::desc("Force top-down list scheduling"));
41 cl::opt<bool> ForceBottomUp("misched-bottomup", cl::Hidden,
42 cl::desc("Force bottom-up list scheduling"));
46 static cl::opt<bool> ViewMISchedDAGs("view-misched-dags", cl::Hidden,
47 cl::desc("Pop up a window to show MISched dags after they are processed"));
49 static cl::opt<unsigned> MISchedCutoff("misched-cutoff", cl::Hidden,
50 cl::desc("Stop scheduling after N instructions"), cl::init(~0U));
52 static bool ViewMISchedDAGs = false;
55 // FIXME: remove this flag after initial testing. It should always be a good
57 static cl::opt<bool> EnableCopyConstrain("misched-vcopy", cl::Hidden,
58 cl::desc("Constrain vreg copies."), cl::init(true));
60 static cl::opt<bool> EnableLoadCluster("misched-cluster", cl::Hidden,
61 cl::desc("Enable load clustering."), cl::init(true));
63 // Experimental heuristics
64 static cl::opt<bool> EnableMacroFusion("misched-fusion", cl::Hidden,
65 cl::desc("Enable scheduling for macro fusion."), cl::init(true));
67 static cl::opt<bool> VerifyScheduling("verify-misched", cl::Hidden,
68 cl::desc("Verify machine instrs before and after machine scheduling"));
70 // DAG subtrees must have at least this many nodes.
71 static const unsigned MinSubtreeSize = 8;
73 //===----------------------------------------------------------------------===//
74 // Machine Instruction Scheduling Pass and Registry
75 //===----------------------------------------------------------------------===//
77 MachineSchedContext::MachineSchedContext():
78 MF(0), MLI(0), MDT(0), PassConfig(0), AA(0), LIS(0) {
79 RegClassInfo = new RegisterClassInfo();
82 MachineSchedContext::~MachineSchedContext() {
87 /// MachineScheduler runs after coalescing and before register allocation.
88 class MachineScheduler : public MachineSchedContext,
89 public MachineFunctionPass {
93 virtual void getAnalysisUsage(AnalysisUsage &AU) const;
95 virtual void releaseMemory() {}
97 virtual bool runOnMachineFunction(MachineFunction&);
99 virtual void print(raw_ostream &O, const Module* = 0) const;
101 static char ID; // Class identification, replacement for typeinfo
105 char MachineScheduler::ID = 0;
107 char &llvm::MachineSchedulerID = MachineScheduler::ID;
109 INITIALIZE_PASS_BEGIN(MachineScheduler, "misched",
110 "Machine Instruction Scheduler", false, false)
111 INITIALIZE_AG_DEPENDENCY(AliasAnalysis)
112 INITIALIZE_PASS_DEPENDENCY(SlotIndexes)
113 INITIALIZE_PASS_DEPENDENCY(LiveIntervals)
114 INITIALIZE_PASS_END(MachineScheduler, "misched",
115 "Machine Instruction Scheduler", false, false)
117 MachineScheduler::MachineScheduler()
118 : MachineFunctionPass(ID) {
119 initializeMachineSchedulerPass(*PassRegistry::getPassRegistry());
122 void MachineScheduler::getAnalysisUsage(AnalysisUsage &AU) const {
123 AU.setPreservesCFG();
124 AU.addRequiredID(MachineDominatorsID);
125 AU.addRequired<MachineLoopInfo>();
126 AU.addRequired<AliasAnalysis>();
127 AU.addRequired<TargetPassConfig>();
128 AU.addRequired<SlotIndexes>();
129 AU.addPreserved<SlotIndexes>();
130 AU.addRequired<LiveIntervals>();
131 AU.addPreserved<LiveIntervals>();
132 MachineFunctionPass::getAnalysisUsage(AU);
135 MachinePassRegistry MachineSchedRegistry::Registry;
137 /// A dummy default scheduler factory indicates whether the scheduler
138 /// is overridden on the command line.
139 static ScheduleDAGInstrs *useDefaultMachineSched(MachineSchedContext *C) {
143 /// MachineSchedOpt allows command line selection of the scheduler.
144 static cl::opt<MachineSchedRegistry::ScheduleDAGCtor, false,
145 RegisterPassParser<MachineSchedRegistry> >
146 MachineSchedOpt("misched",
147 cl::init(&useDefaultMachineSched), cl::Hidden,
148 cl::desc("Machine instruction scheduler to use"));
150 static MachineSchedRegistry
151 DefaultSchedRegistry("default", "Use the target's default scheduler choice.",
152 useDefaultMachineSched);
154 /// Forward declare the standard machine scheduler. This will be used as the
155 /// default scheduler if the target does not set a default.
156 static ScheduleDAGInstrs *createConvergingSched(MachineSchedContext *C);
159 /// Decrement this iterator until reaching the top or a non-debug instr.
160 static MachineBasicBlock::iterator
161 priorNonDebug(MachineBasicBlock::iterator I, MachineBasicBlock::iterator Beg) {
162 assert(I != Beg && "reached the top of the region, cannot decrement");
164 if (!I->isDebugValue())
170 /// If this iterator is a debug value, increment until reaching the End or a
171 /// non-debug instruction.
172 static MachineBasicBlock::iterator
173 nextIfDebug(MachineBasicBlock::iterator I, MachineBasicBlock::iterator End) {
174 for(; I != End; ++I) {
175 if (!I->isDebugValue())
181 /// Top-level MachineScheduler pass driver.
183 /// Visit blocks in function order. Divide each block into scheduling regions
184 /// and visit them bottom-up. Visiting regions bottom-up is not required, but is
185 /// consistent with the DAG builder, which traverses the interior of the
186 /// scheduling regions bottom-up.
188 /// This design avoids exposing scheduling boundaries to the DAG builder,
189 /// simplifying the DAG builder's support for "special" target instructions.
190 /// At the same time the design allows target schedulers to operate across
191 /// scheduling boundaries, for example to bundle the boudary instructions
192 /// without reordering them. This creates complexity, because the target
193 /// scheduler must update the RegionBegin and RegionEnd positions cached by
194 /// ScheduleDAGInstrs whenever adding or removing instructions. A much simpler
195 /// design would be to split blocks at scheduling boundaries, but LLVM has a
196 /// general bias against block splitting purely for implementation simplicity.
197 bool MachineScheduler::runOnMachineFunction(MachineFunction &mf) {
198 DEBUG(dbgs() << "Before MISsched:\n"; mf.print(dbgs()));
200 // Initialize the context of the pass.
202 MLI = &getAnalysis<MachineLoopInfo>();
203 MDT = &getAnalysis<MachineDominatorTree>();
204 PassConfig = &getAnalysis<TargetPassConfig>();
205 AA = &getAnalysis<AliasAnalysis>();
207 LIS = &getAnalysis<LiveIntervals>();
208 const TargetInstrInfo *TII = MF->getTarget().getInstrInfo();
210 if (VerifyScheduling) {
211 DEBUG(LIS->print(dbgs()));
212 MF->verify(this, "Before machine scheduling.");
214 RegClassInfo->runOnMachineFunction(*MF);
216 // Select the scheduler, or set the default.
217 MachineSchedRegistry::ScheduleDAGCtor Ctor = MachineSchedOpt;
218 if (Ctor == useDefaultMachineSched) {
219 // Get the default scheduler set by the target.
220 Ctor = MachineSchedRegistry::getDefault();
222 Ctor = createConvergingSched;
223 MachineSchedRegistry::setDefault(Ctor);
226 // Instantiate the selected scheduler.
227 OwningPtr<ScheduleDAGInstrs> Scheduler(Ctor(this));
229 // Visit all machine basic blocks.
231 // TODO: Visit blocks in global postorder or postorder within the bottom-up
232 // loop tree. Then we can optionally compute global RegPressure.
233 for (MachineFunction::iterator MBB = MF->begin(), MBBEnd = MF->end();
234 MBB != MBBEnd; ++MBB) {
236 Scheduler->startBlock(MBB);
238 // Break the block into scheduling regions [I, RegionEnd), and schedule each
239 // region as soon as it is discovered. RegionEnd points the scheduling
240 // boundary at the bottom of the region. The DAG does not include RegionEnd,
241 // but the region does (i.e. the next RegionEnd is above the previous
242 // RegionBegin). If the current block has no terminator then RegionEnd ==
243 // MBB->end() for the bottom region.
245 // The Scheduler may insert instructions during either schedule() or
246 // exitRegion(), even for empty regions. So the local iterators 'I' and
247 // 'RegionEnd' are invalid across these calls.
248 unsigned RemainingInstrs = MBB->size();
249 for(MachineBasicBlock::iterator RegionEnd = MBB->end();
250 RegionEnd != MBB->begin(); RegionEnd = Scheduler->begin()) {
252 // Avoid decrementing RegionEnd for blocks with no terminator.
253 if (RegionEnd != MBB->end()
254 || TII->isSchedulingBoundary(llvm::prior(RegionEnd), MBB, *MF)) {
256 // Count the boundary instruction.
260 // The next region starts above the previous region. Look backward in the
261 // instruction stream until we find the nearest boundary.
262 MachineBasicBlock::iterator I = RegionEnd;
263 for(;I != MBB->begin(); --I, --RemainingInstrs) {
264 if (TII->isSchedulingBoundary(llvm::prior(I), MBB, *MF))
267 // Notify the scheduler of the region, even if we may skip scheduling
268 // it. Perhaps it still needs to be bundled.
269 Scheduler->enterRegion(MBB, I, RegionEnd, RemainingInstrs);
271 // Skip empty scheduling regions (0 or 1 schedulable instructions).
272 if (I == RegionEnd || I == llvm::prior(RegionEnd)) {
273 // Close the current region. Bundle the terminator if needed.
274 // This invalidates 'RegionEnd' and 'I'.
275 Scheduler->exitRegion();
278 DEBUG(dbgs() << "********** MI Scheduling **********\n");
279 DEBUG(dbgs() << MF->getName()
280 << ":BB#" << MBB->getNumber() << " " << MBB->getName()
281 << "\n From: " << *I << " To: ";
282 if (RegionEnd != MBB->end()) dbgs() << *RegionEnd;
283 else dbgs() << "End";
284 dbgs() << " Remaining: " << RemainingInstrs << "\n");
286 // Schedule a region: possibly reorder instructions.
287 // This invalidates 'RegionEnd' and 'I'.
288 Scheduler->schedule();
290 // Close the current region.
291 Scheduler->exitRegion();
293 // Scheduling has invalidated the current iterator 'I'. Ask the
294 // scheduler for the top of it's scheduled region.
295 RegionEnd = Scheduler->begin();
297 assert(RemainingInstrs == 0 && "Instruction count mismatch!");
298 Scheduler->finishBlock();
300 Scheduler->finalizeSchedule();
301 DEBUG(LIS->print(dbgs()));
302 if (VerifyScheduling)
303 MF->verify(this, "After machine scheduling.");
307 void MachineScheduler::print(raw_ostream &O, const Module* m) const {
311 #if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
312 void ReadyQueue::dump() {
313 dbgs() << " " << Name << ": ";
314 for (unsigned i = 0, e = Queue.size(); i < e; ++i)
315 dbgs() << Queue[i]->NodeNum << " ";
320 //===----------------------------------------------------------------------===//
321 // ScheduleDAGMI - Base class for MachineInstr scheduling with LiveIntervals
323 //===----------------------------------------------------------------------===//
325 ScheduleDAGMI::~ScheduleDAGMI() {
327 DeleteContainerPointers(Mutations);
331 bool ScheduleDAGMI::canAddEdge(SUnit *SuccSU, SUnit *PredSU) {
332 return SuccSU == &ExitSU || !Topo.IsReachable(PredSU, SuccSU);
335 bool ScheduleDAGMI::addEdge(SUnit *SuccSU, const SDep &PredDep) {
336 if (SuccSU != &ExitSU) {
337 // Do not use WillCreateCycle, it assumes SD scheduling.
338 // If Pred is reachable from Succ, then the edge creates a cycle.
339 if (Topo.IsReachable(PredDep.getSUnit(), SuccSU))
341 Topo.AddPred(SuccSU, PredDep.getSUnit());
343 SuccSU->addPred(PredDep, /*Required=*/!PredDep.isArtificial());
344 // Return true regardless of whether a new edge needed to be inserted.
348 /// ReleaseSucc - Decrement the NumPredsLeft count of a successor. When
349 /// NumPredsLeft reaches zero, release the successor node.
351 /// FIXME: Adjust SuccSU height based on MinLatency.
352 void ScheduleDAGMI::releaseSucc(SUnit *SU, SDep *SuccEdge) {
353 SUnit *SuccSU = SuccEdge->getSUnit();
355 if (SuccEdge->isWeak()) {
356 --SuccSU->WeakPredsLeft;
357 if (SuccEdge->isCluster())
358 NextClusterSucc = SuccSU;
362 if (SuccSU->NumPredsLeft == 0) {
363 dbgs() << "*** Scheduling failed! ***\n";
365 dbgs() << " has been released too many times!\n";
369 --SuccSU->NumPredsLeft;
370 if (SuccSU->NumPredsLeft == 0 && SuccSU != &ExitSU)
371 SchedImpl->releaseTopNode(SuccSU);
374 /// releaseSuccessors - Call releaseSucc on each of SU's successors.
375 void ScheduleDAGMI::releaseSuccessors(SUnit *SU) {
376 for (SUnit::succ_iterator I = SU->Succs.begin(), E = SU->Succs.end();
378 releaseSucc(SU, &*I);
382 /// ReleasePred - Decrement the NumSuccsLeft count of a predecessor. When
383 /// NumSuccsLeft reaches zero, release the predecessor node.
385 /// FIXME: Adjust PredSU height based on MinLatency.
386 void ScheduleDAGMI::releasePred(SUnit *SU, SDep *PredEdge) {
387 SUnit *PredSU = PredEdge->getSUnit();
389 if (PredEdge->isWeak()) {
390 --PredSU->WeakSuccsLeft;
391 if (PredEdge->isCluster())
392 NextClusterPred = PredSU;
396 if (PredSU->NumSuccsLeft == 0) {
397 dbgs() << "*** Scheduling failed! ***\n";
399 dbgs() << " has been released too many times!\n";
403 --PredSU->NumSuccsLeft;
404 if (PredSU->NumSuccsLeft == 0 && PredSU != &EntrySU)
405 SchedImpl->releaseBottomNode(PredSU);
408 /// releasePredecessors - Call releasePred on each of SU's predecessors.
409 void ScheduleDAGMI::releasePredecessors(SUnit *SU) {
410 for (SUnit::pred_iterator I = SU->Preds.begin(), E = SU->Preds.end();
412 releasePred(SU, &*I);
416 /// This is normally called from the main scheduler loop but may also be invoked
417 /// by the scheduling strategy to perform additional code motion.
418 void ScheduleDAGMI::moveInstruction(MachineInstr *MI,
419 MachineBasicBlock::iterator InsertPos) {
420 // Advance RegionBegin if the first instruction moves down.
421 if (&*RegionBegin == MI)
424 // Update the instruction stream.
425 BB->splice(InsertPos, BB, MI);
427 // Update LiveIntervals
428 LIS->handleMove(MI, /*UpdateFlags=*/true);
430 // Recede RegionBegin if an instruction moves above the first.
431 if (RegionBegin == InsertPos)
435 bool ScheduleDAGMI::checkSchedLimit() {
437 if (NumInstrsScheduled == MISchedCutoff && MISchedCutoff != ~0U) {
438 CurrentTop = CurrentBottom;
441 ++NumInstrsScheduled;
446 /// enterRegion - Called back from MachineScheduler::runOnMachineFunction after
447 /// crossing a scheduling boundary. [begin, end) includes all instructions in
448 /// the region, including the boundary itself and single-instruction regions
449 /// that don't get scheduled.
450 void ScheduleDAGMI::enterRegion(MachineBasicBlock *bb,
451 MachineBasicBlock::iterator begin,
452 MachineBasicBlock::iterator end,
455 ScheduleDAGInstrs::enterRegion(bb, begin, end, endcount);
457 // For convenience remember the end of the liveness region.
459 (RegionEnd == bb->end()) ? RegionEnd : llvm::next(RegionEnd);
462 // Setup the register pressure trackers for the top scheduled top and bottom
463 // scheduled regions.
464 void ScheduleDAGMI::initRegPressure() {
465 TopRPTracker.init(&MF, RegClassInfo, LIS, BB, RegionBegin);
466 BotRPTracker.init(&MF, RegClassInfo, LIS, BB, LiveRegionEnd);
468 // Close the RPTracker to finalize live ins.
469 RPTracker.closeRegion();
471 DEBUG(RPTracker.getPressure().dump(TRI));
473 // Initialize the live ins and live outs.
474 TopRPTracker.addLiveRegs(RPTracker.getPressure().LiveInRegs);
475 BotRPTracker.addLiveRegs(RPTracker.getPressure().LiveOutRegs);
477 // Close one end of the tracker so we can call
478 // getMaxUpward/DownwardPressureDelta before advancing across any
479 // instructions. This converts currently live regs into live ins/outs.
480 TopRPTracker.closeTop();
481 BotRPTracker.closeBottom();
483 // Account for liveness generated by the region boundary.
484 if (LiveRegionEnd != RegionEnd)
485 BotRPTracker.recede();
487 assert(BotRPTracker.getPos() == RegionEnd && "Can't find the region bottom");
489 // Cache the list of excess pressure sets in this region. This will also track
490 // the max pressure in the scheduled code for these sets.
491 RegionCriticalPSets.clear();
492 const std::vector<unsigned> &RegionPressure =
493 RPTracker.getPressure().MaxSetPressure;
494 for (unsigned i = 0, e = RegionPressure.size(); i < e; ++i) {
495 unsigned Limit = TRI->getRegPressureSetLimit(i);
496 DEBUG(dbgs() << TRI->getRegPressureSetName(i)
498 << " Actual " << RegionPressure[i] << "\n");
499 if (RegionPressure[i] > Limit)
500 RegionCriticalPSets.push_back(PressureElement(i, 0));
502 DEBUG(dbgs() << "Excess PSets: ";
503 for (unsigned i = 0, e = RegionCriticalPSets.size(); i != e; ++i)
504 dbgs() << TRI->getRegPressureSetName(
505 RegionCriticalPSets[i].PSetID) << " ";
509 // FIXME: When the pressure tracker deals in pressure differences then we won't
510 // iterate over all RegionCriticalPSets[i].
512 updateScheduledPressure(const std::vector<unsigned> &NewMaxPressure) {
513 for (unsigned i = 0, e = RegionCriticalPSets.size(); i < e; ++i) {
514 unsigned ID = RegionCriticalPSets[i].PSetID;
515 int &MaxUnits = RegionCriticalPSets[i].UnitIncrease;
516 if ((int)NewMaxPressure[ID] > MaxUnits)
517 MaxUnits = NewMaxPressure[ID];
520 for (unsigned i = 0, e = NewMaxPressure.size(); i < e; ++i) {
521 unsigned Limit = TRI->getRegPressureSetLimit(i);
522 if (NewMaxPressure[i] > Limit ) {
523 dbgs() << " " << TRI->getRegPressureSetName(i) << ": "
524 << NewMaxPressure[i] << " > " << Limit << "\n";
529 /// schedule - Called back from MachineScheduler::runOnMachineFunction
530 /// after setting up the current scheduling region. [RegionBegin, RegionEnd)
531 /// only includes instructions that have DAG nodes, not scheduling boundaries.
533 /// This is a skeletal driver, with all the functionality pushed into helpers,
534 /// so that it can be easilly extended by experimental schedulers. Generally,
535 /// implementing MachineSchedStrategy should be sufficient to implement a new
536 /// scheduling algorithm. However, if a scheduler further subclasses
537 /// ScheduleDAGMI then it will want to override this virtual method in order to
538 /// update any specialized state.
539 void ScheduleDAGMI::schedule() {
540 buildDAGWithRegPressure();
542 Topo.InitDAGTopologicalSorting();
546 SmallVector<SUnit*, 8> TopRoots, BotRoots;
547 findRootsAndBiasEdges(TopRoots, BotRoots);
549 // Initialize the strategy before modifying the DAG.
550 // This may initialize a DFSResult to be used for queue priority.
551 SchedImpl->initialize(this);
553 DEBUG(for (unsigned su = 0, e = SUnits.size(); su != e; ++su)
554 SUnits[su].dumpAll(this));
555 if (ViewMISchedDAGs) viewGraph();
557 // Initialize ready queues now that the DAG and priority data are finalized.
558 initQueues(TopRoots, BotRoots);
560 bool IsTopNode = false;
561 while (SUnit *SU = SchedImpl->pickNode(IsTopNode)) {
562 assert(!SU->isScheduled && "Node already scheduled");
563 if (!checkSchedLimit())
566 scheduleMI(SU, IsTopNode);
568 updateQueues(SU, IsTopNode);
570 assert(CurrentTop == CurrentBottom && "Nonempty unscheduled zone.");
575 unsigned BBNum = begin()->getParent()->getNumber();
576 dbgs() << "*** Final schedule for BB#" << BBNum << " ***\n";
582 /// Build the DAG and setup three register pressure trackers.
583 void ScheduleDAGMI::buildDAGWithRegPressure() {
584 // Initialize the register pressure tracker used by buildSchedGraph.
585 RPTracker.init(&MF, RegClassInfo, LIS, BB, LiveRegionEnd);
587 // Account for liveness generate by the region boundary.
588 if (LiveRegionEnd != RegionEnd)
591 // Build the DAG, and compute current register pressure.
592 buildSchedGraph(AA, &RPTracker);
594 // Initialize top/bottom trackers after computing region pressure.
598 /// Apply each ScheduleDAGMutation step in order.
599 void ScheduleDAGMI::postprocessDAG() {
600 for (unsigned i = 0, e = Mutations.size(); i < e; ++i) {
601 Mutations[i]->apply(this);
605 void ScheduleDAGMI::computeDFSResult() {
607 DFSResult = new SchedDFSResult(/*BottomU*/true, MinSubtreeSize);
609 ScheduledTrees.clear();
610 DFSResult->resize(SUnits.size());
611 DFSResult->compute(SUnits);
612 ScheduledTrees.resize(DFSResult->getNumSubtrees());
615 void ScheduleDAGMI::findRootsAndBiasEdges(SmallVectorImpl<SUnit*> &TopRoots,
616 SmallVectorImpl<SUnit*> &BotRoots) {
617 for (std::vector<SUnit>::iterator
618 I = SUnits.begin(), E = SUnits.end(); I != E; ++I) {
620 assert(!SU->isBoundaryNode() && "Boundary node should not be in SUnits");
622 // Order predecessors so DFSResult follows the critical path.
623 SU->biasCriticalPath();
625 // A SUnit is ready to top schedule if it has no predecessors.
626 if (!I->NumPredsLeft)
627 TopRoots.push_back(SU);
628 // A SUnit is ready to bottom schedule if it has no successors.
629 if (!I->NumSuccsLeft)
630 BotRoots.push_back(SU);
632 ExitSU.biasCriticalPath();
635 /// Identify DAG roots and setup scheduler queues.
636 void ScheduleDAGMI::initQueues(ArrayRef<SUnit*> TopRoots,
637 ArrayRef<SUnit*> BotRoots) {
638 NextClusterSucc = NULL;
639 NextClusterPred = NULL;
641 // Release all DAG roots for scheduling, not including EntrySU/ExitSU.
643 // Nodes with unreleased weak edges can still be roots.
644 // Release top roots in forward order.
645 for (SmallVectorImpl<SUnit*>::const_iterator
646 I = TopRoots.begin(), E = TopRoots.end(); I != E; ++I) {
647 SchedImpl->releaseTopNode(*I);
649 // Release bottom roots in reverse order so the higher priority nodes appear
650 // first. This is more natural and slightly more efficient.
651 for (SmallVectorImpl<SUnit*>::const_reverse_iterator
652 I = BotRoots.rbegin(), E = BotRoots.rend(); I != E; ++I) {
653 SchedImpl->releaseBottomNode(*I);
656 releaseSuccessors(&EntrySU);
657 releasePredecessors(&ExitSU);
659 SchedImpl->registerRoots();
661 // Advance past initial DebugValues.
662 assert(TopRPTracker.getPos() == RegionBegin && "bad initial Top tracker");
663 CurrentTop = nextIfDebug(RegionBegin, RegionEnd);
664 TopRPTracker.setPos(CurrentTop);
666 CurrentBottom = RegionEnd;
669 /// Move an instruction and update register pressure.
670 void ScheduleDAGMI::scheduleMI(SUnit *SU, bool IsTopNode) {
671 // Move the instruction to its new location in the instruction stream.
672 MachineInstr *MI = SU->getInstr();
675 assert(SU->isTopReady() && "node still has unscheduled dependencies");
676 if (&*CurrentTop == MI)
677 CurrentTop = nextIfDebug(++CurrentTop, CurrentBottom);
679 moveInstruction(MI, CurrentTop);
680 TopRPTracker.setPos(MI);
683 // Update top scheduled pressure.
684 TopRPTracker.advance();
685 assert(TopRPTracker.getPos() == CurrentTop && "out of sync");
686 updateScheduledPressure(TopRPTracker.getPressure().MaxSetPressure);
689 assert(SU->isBottomReady() && "node still has unscheduled dependencies");
690 MachineBasicBlock::iterator priorII =
691 priorNonDebug(CurrentBottom, CurrentTop);
693 CurrentBottom = priorII;
695 if (&*CurrentTop == MI) {
696 CurrentTop = nextIfDebug(++CurrentTop, priorII);
697 TopRPTracker.setPos(CurrentTop);
699 moveInstruction(MI, CurrentBottom);
702 // Update bottom scheduled pressure.
703 BotRPTracker.recede();
704 assert(BotRPTracker.getPos() == CurrentBottom && "out of sync");
705 updateScheduledPressure(BotRPTracker.getPressure().MaxSetPressure);
709 /// Update scheduler queues after scheduling an instruction.
710 void ScheduleDAGMI::updateQueues(SUnit *SU, bool IsTopNode) {
711 // Release dependent instructions for scheduling.
713 releaseSuccessors(SU);
715 releasePredecessors(SU);
717 SU->isScheduled = true;
720 unsigned SubtreeID = DFSResult->getSubtreeID(SU);
721 if (!ScheduledTrees.test(SubtreeID)) {
722 ScheduledTrees.set(SubtreeID);
723 DFSResult->scheduleTree(SubtreeID);
724 SchedImpl->scheduleTree(SubtreeID);
728 // Notify the scheduling strategy after updating the DAG.
729 SchedImpl->schedNode(SU, IsTopNode);
732 /// Reinsert any remaining debug_values, just like the PostRA scheduler.
733 void ScheduleDAGMI::placeDebugValues() {
734 // If first instruction was a DBG_VALUE then put it back.
736 BB->splice(RegionBegin, BB, FirstDbgValue);
737 RegionBegin = FirstDbgValue;
740 for (std::vector<std::pair<MachineInstr *, MachineInstr *> >::iterator
741 DI = DbgValues.end(), DE = DbgValues.begin(); DI != DE; --DI) {
742 std::pair<MachineInstr *, MachineInstr *> P = *prior(DI);
743 MachineInstr *DbgValue = P.first;
744 MachineBasicBlock::iterator OrigPrevMI = P.second;
745 if (&*RegionBegin == DbgValue)
747 BB->splice(++OrigPrevMI, BB, DbgValue);
748 if (OrigPrevMI == llvm::prior(RegionEnd))
749 RegionEnd = DbgValue;
752 FirstDbgValue = NULL;
755 #if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
756 void ScheduleDAGMI::dumpSchedule() const {
757 for (MachineBasicBlock::iterator MI = begin(), ME = end(); MI != ME; ++MI) {
758 if (SUnit *SU = getSUnit(&(*MI)))
761 dbgs() << "Missing SUnit\n";
766 //===----------------------------------------------------------------------===//
767 // LoadClusterMutation - DAG post-processing to cluster loads.
768 //===----------------------------------------------------------------------===//
771 /// \brief Post-process the DAG to create cluster edges between neighboring
773 class LoadClusterMutation : public ScheduleDAGMutation {
778 LoadInfo(SUnit *su, unsigned reg, unsigned ofs)
779 : SU(su), BaseReg(reg), Offset(ofs) {}
781 static bool LoadInfoLess(const LoadClusterMutation::LoadInfo &LHS,
782 const LoadClusterMutation::LoadInfo &RHS);
784 const TargetInstrInfo *TII;
785 const TargetRegisterInfo *TRI;
787 LoadClusterMutation(const TargetInstrInfo *tii,
788 const TargetRegisterInfo *tri)
789 : TII(tii), TRI(tri) {}
791 virtual void apply(ScheduleDAGMI *DAG);
793 void clusterNeighboringLoads(ArrayRef<SUnit*> Loads, ScheduleDAGMI *DAG);
797 bool LoadClusterMutation::LoadInfoLess(
798 const LoadClusterMutation::LoadInfo &LHS,
799 const LoadClusterMutation::LoadInfo &RHS) {
800 if (LHS.BaseReg != RHS.BaseReg)
801 return LHS.BaseReg < RHS.BaseReg;
802 return LHS.Offset < RHS.Offset;
805 void LoadClusterMutation::clusterNeighboringLoads(ArrayRef<SUnit*> Loads,
806 ScheduleDAGMI *DAG) {
807 SmallVector<LoadClusterMutation::LoadInfo,32> LoadRecords;
808 for (unsigned Idx = 0, End = Loads.size(); Idx != End; ++Idx) {
809 SUnit *SU = Loads[Idx];
812 if (TII->getLdStBaseRegImmOfs(SU->getInstr(), BaseReg, Offset, TRI))
813 LoadRecords.push_back(LoadInfo(SU, BaseReg, Offset));
815 if (LoadRecords.size() < 2)
817 std::sort(LoadRecords.begin(), LoadRecords.end(), LoadInfoLess);
818 unsigned ClusterLength = 1;
819 for (unsigned Idx = 0, End = LoadRecords.size(); Idx < (End - 1); ++Idx) {
820 if (LoadRecords[Idx].BaseReg != LoadRecords[Idx+1].BaseReg) {
825 SUnit *SUa = LoadRecords[Idx].SU;
826 SUnit *SUb = LoadRecords[Idx+1].SU;
827 if (TII->shouldClusterLoads(SUa->getInstr(), SUb->getInstr(), ClusterLength)
828 && DAG->addEdge(SUb, SDep(SUa, SDep::Cluster))) {
830 DEBUG(dbgs() << "Cluster loads SU(" << SUa->NodeNum << ") - SU("
831 << SUb->NodeNum << ")\n");
832 // Copy successor edges from SUa to SUb. Interleaving computation
833 // dependent on SUa can prevent load combining due to register reuse.
834 // Predecessor edges do not need to be copied from SUb to SUa since nearby
835 // loads should have effectively the same inputs.
836 for (SUnit::const_succ_iterator
837 SI = SUa->Succs.begin(), SE = SUa->Succs.end(); SI != SE; ++SI) {
838 if (SI->getSUnit() == SUb)
840 DEBUG(dbgs() << " Copy Succ SU(" << SI->getSUnit()->NodeNum << ")\n");
841 DAG->addEdge(SI->getSUnit(), SDep(SUb, SDep::Artificial));
850 /// \brief Callback from DAG postProcessing to create cluster edges for loads.
851 void LoadClusterMutation::apply(ScheduleDAGMI *DAG) {
852 // Map DAG NodeNum to store chain ID.
853 DenseMap<unsigned, unsigned> StoreChainIDs;
854 // Map each store chain to a set of dependent loads.
855 SmallVector<SmallVector<SUnit*,4>, 32> StoreChainDependents;
856 for (unsigned Idx = 0, End = DAG->SUnits.size(); Idx != End; ++Idx) {
857 SUnit *SU = &DAG->SUnits[Idx];
858 if (!SU->getInstr()->mayLoad())
860 unsigned ChainPredID = DAG->SUnits.size();
861 for (SUnit::const_pred_iterator
862 PI = SU->Preds.begin(), PE = SU->Preds.end(); PI != PE; ++PI) {
864 ChainPredID = PI->getSUnit()->NodeNum;
868 // Check if this chain-like pred has been seen
869 // before. ChainPredID==MaxNodeID for loads at the top of the schedule.
870 unsigned NumChains = StoreChainDependents.size();
871 std::pair<DenseMap<unsigned, unsigned>::iterator, bool> Result =
872 StoreChainIDs.insert(std::make_pair(ChainPredID, NumChains));
874 StoreChainDependents.resize(NumChains + 1);
875 StoreChainDependents[Result.first->second].push_back(SU);
877 // Iterate over the store chains.
878 for (unsigned Idx = 0, End = StoreChainDependents.size(); Idx != End; ++Idx)
879 clusterNeighboringLoads(StoreChainDependents[Idx], DAG);
882 //===----------------------------------------------------------------------===//
883 // MacroFusion - DAG post-processing to encourage fusion of macro ops.
884 //===----------------------------------------------------------------------===//
887 /// \brief Post-process the DAG to create cluster edges between instructions
888 /// that may be fused by the processor into a single operation.
889 class MacroFusion : public ScheduleDAGMutation {
890 const TargetInstrInfo *TII;
892 MacroFusion(const TargetInstrInfo *tii): TII(tii) {}
894 virtual void apply(ScheduleDAGMI *DAG);
898 /// \brief Callback from DAG postProcessing to create cluster edges to encourage
899 /// fused operations.
900 void MacroFusion::apply(ScheduleDAGMI *DAG) {
901 // For now, assume targets can only fuse with the branch.
902 MachineInstr *Branch = DAG->ExitSU.getInstr();
906 for (unsigned Idx = DAG->SUnits.size(); Idx > 0;) {
907 SUnit *SU = &DAG->SUnits[--Idx];
908 if (!TII->shouldScheduleAdjacent(SU->getInstr(), Branch))
911 // Create a single weak edge from SU to ExitSU. The only effect is to cause
912 // bottom-up scheduling to heavily prioritize the clustered SU. There is no
913 // need to copy predecessor edges from ExitSU to SU, since top-down
914 // scheduling cannot prioritize ExitSU anyway. To defer top-down scheduling
915 // of SU, we could create an artificial edge from the deepest root, but it
916 // hasn't been needed yet.
917 bool Success = DAG->addEdge(&DAG->ExitSU, SDep(SU, SDep::Cluster));
919 assert(Success && "No DAG nodes should be reachable from ExitSU");
921 DEBUG(dbgs() << "Macro Fuse SU(" << SU->NodeNum << ")\n");
926 //===----------------------------------------------------------------------===//
927 // CopyConstrain - DAG post-processing to encourage copy elimination.
928 //===----------------------------------------------------------------------===//
931 /// \brief Post-process the DAG to create weak edges from all uses of a copy to
932 /// the one use that defines the copy's source vreg, most likely an induction
933 /// variable increment.
934 class CopyConstrain : public ScheduleDAGMutation {
936 SlotIndex RegionBeginIdx;
937 // RegionEndIdx is the slot index of the last non-debug instruction in the
938 // scheduling region. So we may have RegionBeginIdx == RegionEndIdx.
939 SlotIndex RegionEndIdx;
941 CopyConstrain(const TargetInstrInfo *, const TargetRegisterInfo *) {}
943 virtual void apply(ScheduleDAGMI *DAG);
946 void constrainLocalCopy(SUnit *CopySU, ScheduleDAGMI *DAG);
950 /// constrainLocalCopy handles two possibilities:
955 /// I3: dst = src (copy)
956 /// (create pred->succ edges I0->I1, I2->I1)
959 /// I0: dst = src (copy)
963 /// (create pred->succ edges I1->I2, I3->I2)
965 /// Although the MachineScheduler is currently constrained to single blocks,
966 /// this algorithm should handle extended blocks. An EBB is a set of
967 /// contiguously numbered blocks such that the previous block in the EBB is
968 /// always the single predecessor.
969 void CopyConstrain::constrainLocalCopy(SUnit *CopySU, ScheduleDAGMI *DAG) {
970 LiveIntervals *LIS = DAG->getLIS();
971 MachineInstr *Copy = CopySU->getInstr();
973 // Check for pure vreg copies.
974 unsigned SrcReg = Copy->getOperand(1).getReg();
975 if (!TargetRegisterInfo::isVirtualRegister(SrcReg))
978 unsigned DstReg = Copy->getOperand(0).getReg();
979 if (!TargetRegisterInfo::isVirtualRegister(DstReg))
982 // Check if either the dest or source is local. If it's live across a back
983 // edge, it's not local. Note that if both vregs are live across the back
984 // edge, we cannot successfully contrain the copy without cyclic scheduling.
985 unsigned LocalReg = DstReg;
986 unsigned GlobalReg = SrcReg;
987 LiveInterval *LocalLI = &LIS->getInterval(LocalReg);
988 if (!LocalLI->isLocal(RegionBeginIdx, RegionEndIdx)) {
991 LocalLI = &LIS->getInterval(LocalReg);
992 if (!LocalLI->isLocal(RegionBeginIdx, RegionEndIdx))
995 LiveInterval *GlobalLI = &LIS->getInterval(GlobalReg);
997 // Find the global segment after the start of the local LI.
998 LiveInterval::iterator GlobalSegment = GlobalLI->find(LocalLI->beginIndex());
999 // If GlobalLI does not overlap LocalLI->start, then a copy directly feeds a
1000 // local live range. We could create edges from other global uses to the local
1001 // start, but the coalescer should have already eliminated these cases, so
1002 // don't bother dealing with it.
1003 if (GlobalSegment == GlobalLI->end())
1006 // If GlobalSegment is killed at the LocalLI->start, the call to find()
1007 // returned the next global segment. But if GlobalSegment overlaps with
1008 // LocalLI->start, then advance to the next segement. If a hole in GlobalLI
1009 // exists in LocalLI's vicinity, GlobalSegment will be the end of the hole.
1010 if (GlobalSegment->contains(LocalLI->beginIndex()))
1013 if (GlobalSegment == GlobalLI->end())
1016 // Check if GlobalLI contains a hole in the vicinity of LocalLI.
1017 if (GlobalSegment != GlobalLI->begin()) {
1018 // Two address defs have no hole.
1019 if (SlotIndex::isSameInstr(llvm::prior(GlobalSegment)->end,
1020 GlobalSegment->start)) {
1023 // If GlobalLI has a prior segment, it must be live into the EBB. Otherwise
1024 // it would be a disconnected component in the live range.
1025 assert(llvm::prior(GlobalSegment)->start < LocalLI->beginIndex() &&
1026 "Disconnected LRG within the scheduling region.");
1028 MachineInstr *GlobalDef = LIS->getInstructionFromIndex(GlobalSegment->start);
1032 SUnit *GlobalSU = DAG->getSUnit(GlobalDef);
1036 // GlobalDef is the bottom of the GlobalLI hole. Open the hole by
1037 // constraining the uses of the last local def to precede GlobalDef.
1038 SmallVector<SUnit*,8> LocalUses;
1039 const VNInfo *LastLocalVN = LocalLI->getVNInfoBefore(LocalLI->endIndex());
1040 MachineInstr *LastLocalDef = LIS->getInstructionFromIndex(LastLocalVN->def);
1041 SUnit *LastLocalSU = DAG->getSUnit(LastLocalDef);
1042 for (SUnit::const_succ_iterator
1043 I = LastLocalSU->Succs.begin(), E = LastLocalSU->Succs.end();
1045 if (I->getKind() != SDep::Data || I->getReg() != LocalReg)
1047 if (I->getSUnit() == GlobalSU)
1049 if (!DAG->canAddEdge(GlobalSU, I->getSUnit()))
1051 LocalUses.push_back(I->getSUnit());
1053 // Open the top of the GlobalLI hole by constraining any earlier global uses
1054 // to precede the start of LocalLI.
1055 SmallVector<SUnit*,8> GlobalUses;
1056 MachineInstr *FirstLocalDef =
1057 LIS->getInstructionFromIndex(LocalLI->beginIndex());
1058 SUnit *FirstLocalSU = DAG->getSUnit(FirstLocalDef);
1059 for (SUnit::const_pred_iterator
1060 I = GlobalSU->Preds.begin(), E = GlobalSU->Preds.end(); I != E; ++I) {
1061 if (I->getKind() != SDep::Anti || I->getReg() != GlobalReg)
1063 if (I->getSUnit() == FirstLocalSU)
1065 if (!DAG->canAddEdge(FirstLocalSU, I->getSUnit()))
1067 GlobalUses.push_back(I->getSUnit());
1069 DEBUG(dbgs() << "Constraining copy SU(" << CopySU->NodeNum << ")\n");
1070 // Add the weak edges.
1071 for (SmallVectorImpl<SUnit*>::const_iterator
1072 I = LocalUses.begin(), E = LocalUses.end(); I != E; ++I) {
1073 DEBUG(dbgs() << " Local use SU(" << (*I)->NodeNum << ") -> SU("
1074 << GlobalSU->NodeNum << ")\n");
1075 DAG->addEdge(GlobalSU, SDep(*I, SDep::Weak));
1077 for (SmallVectorImpl<SUnit*>::const_iterator
1078 I = GlobalUses.begin(), E = GlobalUses.end(); I != E; ++I) {
1079 DEBUG(dbgs() << " Global use SU(" << (*I)->NodeNum << ") -> SU("
1080 << FirstLocalSU->NodeNum << ")\n");
1081 DAG->addEdge(FirstLocalSU, SDep(*I, SDep::Weak));
1085 /// \brief Callback from DAG postProcessing to create weak edges to encourage
1086 /// copy elimination.
1087 void CopyConstrain::apply(ScheduleDAGMI *DAG) {
1088 MachineBasicBlock::iterator FirstPos = nextIfDebug(DAG->begin(), DAG->end());
1089 if (FirstPos == DAG->end())
1091 RegionBeginIdx = DAG->getLIS()->getInstructionIndex(&*FirstPos);
1092 RegionEndIdx = DAG->getLIS()->getInstructionIndex(
1093 &*priorNonDebug(DAG->end(), DAG->begin()));
1095 for (unsigned Idx = 0, End = DAG->SUnits.size(); Idx != End; ++Idx) {
1096 SUnit *SU = &DAG->SUnits[Idx];
1097 if (!SU->getInstr()->isCopy())
1100 constrainLocalCopy(SU, DAG);
1104 //===----------------------------------------------------------------------===//
1105 // ConvergingScheduler - Implementation of the standard MachineSchedStrategy.
1106 //===----------------------------------------------------------------------===//
1109 /// ConvergingScheduler shrinks the unscheduled zone using heuristics to balance
1111 class ConvergingScheduler : public MachineSchedStrategy {
1113 /// Represent the type of SchedCandidate found within a single queue.
1114 /// pickNodeBidirectional depends on these listed by decreasing priority.
1116 NoCand, PhysRegCopy, SingleExcess, SingleCritical, Cluster, Weak,
1117 ResourceReduce, ResourceDemand, BotHeightReduce, BotPathReduce,
1118 TopDepthReduce, TopPathReduce, SingleMax, MultiPressure, NextDefUse,
1122 static const char *getReasonStr(ConvergingScheduler::CandReason Reason);
1125 /// Policy for scheduling the next instruction in the candidate's zone.
1128 unsigned ReduceResIdx;
1129 unsigned DemandResIdx;
1131 CandPolicy(): ReduceLatency(false), ReduceResIdx(0), DemandResIdx(0) {}
1134 /// Status of an instruction's critical resource consumption.
1135 struct SchedResourceDelta {
1136 // Count critical resources in the scheduled region required by SU.
1137 unsigned CritResources;
1139 // Count critical resources from another region consumed by SU.
1140 unsigned DemandedResources;
1142 SchedResourceDelta(): CritResources(0), DemandedResources(0) {}
1144 bool operator==(const SchedResourceDelta &RHS) const {
1145 return CritResources == RHS.CritResources
1146 && DemandedResources == RHS.DemandedResources;
1148 bool operator!=(const SchedResourceDelta &RHS) const {
1149 return !operator==(RHS);
1153 /// Store the state used by ConvergingScheduler heuristics, required for the
1154 /// lifetime of one invocation of pickNode().
1155 struct SchedCandidate {
1158 // The best SUnit candidate.
1161 // The reason for this candidate.
1164 // Register pressure values for the best candidate.
1165 RegPressureDelta RPDelta;
1167 // Critical resource consumption of the best candidate.
1168 SchedResourceDelta ResDelta;
1170 SchedCandidate(const CandPolicy &policy)
1171 : Policy(policy), SU(NULL), Reason(NoCand) {}
1173 bool isValid() const { return SU; }
1175 // Copy the status of another candidate without changing policy.
1176 void setBest(SchedCandidate &Best) {
1177 assert(Best.Reason != NoCand && "uninitialized Sched candidate");
1179 Reason = Best.Reason;
1180 RPDelta = Best.RPDelta;
1181 ResDelta = Best.ResDelta;
1184 void initResourceDelta(const ScheduleDAGMI *DAG,
1185 const TargetSchedModel *SchedModel);
1188 /// Summarize the unscheduled region.
1189 struct SchedRemainder {
1190 // Critical path through the DAG in expected latency.
1191 unsigned CriticalPath;
1193 // Unscheduled resources
1194 SmallVector<unsigned, 16> RemainingCounts;
1195 // Critical resource for the unscheduled zone.
1196 unsigned CritResIdx;
1197 // Number of micro-ops left to schedule.
1198 unsigned RemainingMicroOps;
1202 RemainingCounts.clear();
1204 RemainingMicroOps = 0;
1207 SchedRemainder() { reset(); }
1209 void init(ScheduleDAGMI *DAG, const TargetSchedModel *SchedModel);
1211 unsigned getMaxRemainingCount(const TargetSchedModel *SchedModel) const {
1212 if (!SchedModel->hasInstrSchedModel())
1216 RemainingMicroOps * SchedModel->getMicroOpFactor(),
1217 RemainingCounts[CritResIdx]);
1221 /// Each Scheduling boundary is associated with ready queues. It tracks the
1222 /// current cycle in the direction of movement, and maintains the state
1223 /// of "hazards" and other interlocks at the current cycle.
1224 struct SchedBoundary {
1226 const TargetSchedModel *SchedModel;
1227 SchedRemainder *Rem;
1229 ReadyQueue Available;
1233 // For heuristics, keep a list of the nodes that immediately depend on the
1234 // most recently scheduled node.
1235 SmallPtrSet<const SUnit*, 8> NextSUs;
1237 ScheduleHazardRecognizer *HazardRec;
1240 unsigned IssueCount;
1242 /// MinReadyCycle - Cycle of the soonest available instruction.
1243 unsigned MinReadyCycle;
1245 // The expected latency of the critical path in this scheduled zone.
1246 unsigned ExpectedLatency;
1248 // The latency of dependence chains leading into this zone.
1249 // For each node scheduled: DLat = max DLat, N.Depth.
1250 // For each cycle scheduled: DLat -= 1.
1251 unsigned DependentLatency;
1253 // Resources used in the scheduled zone beyond this boundary.
1254 SmallVector<unsigned, 16> ResourceCounts;
1256 // Cache the critical resources ID in this scheduled zone.
1257 unsigned CritResIdx;
1259 // Is the scheduled region resource limited vs. latency limited.
1260 bool IsResourceLimited;
1262 unsigned ExpectedCount;
1265 // Remember the greatest min operand latency.
1266 unsigned MaxMinLatency;
1270 // A new HazardRec is created for each DAG and owned by SchedBoundary.
1275 CheckPending = false;
1280 MinReadyCycle = UINT_MAX;
1281 ExpectedLatency = 0;
1282 DependentLatency = 0;
1283 ResourceCounts.resize(1);
1284 assert(!ResourceCounts[0] && "nonzero count for bad resource");
1286 IsResourceLimited = false;
1291 // Reserve a zero-count for invalid CritResIdx.
1292 ResourceCounts.resize(1);
1295 /// Pending queues extend the ready queues with the same ID and the
1296 /// PendingFlag set.
1297 SchedBoundary(unsigned ID, const Twine &Name):
1298 DAG(0), SchedModel(0), Rem(0), Available(ID, Name+".A"),
1299 Pending(ID << ConvergingScheduler::LogMaxQID, Name+".P"),
1304 ~SchedBoundary() { delete HazardRec; }
1306 void init(ScheduleDAGMI *dag, const TargetSchedModel *smodel,
1307 SchedRemainder *rem);
1309 bool isTop() const {
1310 return Available.getID() == ConvergingScheduler::TopQID;
1313 unsigned getUnscheduledLatency(SUnit *SU) const {
1315 return SU->getHeight();
1316 return SU->getDepth() + SU->Latency;
1319 unsigned getCriticalCount() const {
1320 return ResourceCounts[CritResIdx];
1323 bool checkHazard(SUnit *SU);
1325 void setLatencyPolicy(CandPolicy &Policy);
1327 void releaseNode(SUnit *SU, unsigned ReadyCycle);
1331 void countResource(unsigned PIdx, unsigned Cycles);
1333 void bumpNode(SUnit *SU);
1335 void releasePending();
1337 void removeReady(SUnit *SU);
1339 SUnit *pickOnlyChoice();
1344 const TargetSchedModel *SchedModel;
1345 const TargetRegisterInfo *TRI;
1347 // State of the top and bottom scheduled instruction boundaries.
1353 /// SUnit::NodeQueueId: 0 (none), 1 (top), 2 (bot), 3 (both)
1360 ConvergingScheduler():
1361 DAG(0), SchedModel(0), TRI(0), Top(TopQID, "TopQ"), Bot(BotQID, "BotQ") {}
1363 virtual void initialize(ScheduleDAGMI *dag);
1365 virtual SUnit *pickNode(bool &IsTopNode);
1367 virtual void schedNode(SUnit *SU, bool IsTopNode);
1369 virtual void releaseTopNode(SUnit *SU);
1371 virtual void releaseBottomNode(SUnit *SU);
1373 virtual void registerRoots();
1377 ConvergingScheduler::SchedBoundary &CriticalZone,
1378 ConvergingScheduler::SchedCandidate &CriticalCand,
1379 ConvergingScheduler::SchedBoundary &OppositeZone,
1380 ConvergingScheduler::SchedCandidate &OppositeCand);
1382 void checkResourceLimits(ConvergingScheduler::SchedCandidate &TopCand,
1383 ConvergingScheduler::SchedCandidate &BotCand);
1385 void tryCandidate(SchedCandidate &Cand,
1386 SchedCandidate &TryCand,
1387 SchedBoundary &Zone,
1388 const RegPressureTracker &RPTracker,
1389 RegPressureTracker &TempTracker);
1391 SUnit *pickNodeBidirectional(bool &IsTopNode);
1393 void pickNodeFromQueue(SchedBoundary &Zone,
1394 const RegPressureTracker &RPTracker,
1395 SchedCandidate &Candidate);
1397 void reschedulePhysRegCopies(SUnit *SU, bool isTop);
1400 void traceCandidate(const SchedCandidate &Cand);
1405 void ConvergingScheduler::SchedRemainder::
1406 init(ScheduleDAGMI *DAG, const TargetSchedModel *SchedModel) {
1408 if (!SchedModel->hasInstrSchedModel())
1410 RemainingCounts.resize(SchedModel->getNumProcResourceKinds());
1411 for (std::vector<SUnit>::iterator
1412 I = DAG->SUnits.begin(), E = DAG->SUnits.end(); I != E; ++I) {
1413 const MCSchedClassDesc *SC = DAG->getSchedClass(&*I);
1414 RemainingMicroOps += SchedModel->getNumMicroOps(I->getInstr(), SC);
1415 for (TargetSchedModel::ProcResIter
1416 PI = SchedModel->getWriteProcResBegin(SC),
1417 PE = SchedModel->getWriteProcResEnd(SC); PI != PE; ++PI) {
1418 unsigned PIdx = PI->ProcResourceIdx;
1419 unsigned Factor = SchedModel->getResourceFactor(PIdx);
1420 RemainingCounts[PIdx] += (Factor * PI->Cycles);
1423 for (unsigned PIdx = 0, PEnd = SchedModel->getNumProcResourceKinds();
1424 PIdx != PEnd; ++PIdx) {
1425 if ((int)(RemainingCounts[PIdx] - RemainingCounts[CritResIdx])
1426 >= (int)SchedModel->getLatencyFactor()) {
1430 DEBUG(dbgs() << "Critical Resource: "
1431 << SchedModel->getProcResource(CritResIdx)->Name
1432 << ": " << RemainingCounts[CritResIdx]
1433 << " / " << SchedModel->getLatencyFactor() << '\n');
1436 void ConvergingScheduler::SchedBoundary::
1437 init(ScheduleDAGMI *dag, const TargetSchedModel *smodel, SchedRemainder *rem) {
1440 SchedModel = smodel;
1442 if (SchedModel->hasInstrSchedModel())
1443 ResourceCounts.resize(SchedModel->getNumProcResourceKinds());
1446 void ConvergingScheduler::initialize(ScheduleDAGMI *dag) {
1448 SchedModel = DAG->getSchedModel();
1451 Rem.init(DAG, SchedModel);
1452 Top.init(DAG, SchedModel, &Rem);
1453 Bot.init(DAG, SchedModel, &Rem);
1455 // Initialize resource counts.
1457 // Initialize the HazardRecognizers. If itineraries don't exist, are empty, or
1458 // are disabled, then these HazardRecs will be disabled.
1459 const InstrItineraryData *Itin = SchedModel->getInstrItineraries();
1460 const TargetMachine &TM = DAG->MF.getTarget();
1461 Top.HazardRec = TM.getInstrInfo()->CreateTargetMIHazardRecognizer(Itin, DAG);
1462 Bot.HazardRec = TM.getInstrInfo()->CreateTargetMIHazardRecognizer(Itin, DAG);
1464 assert((!ForceTopDown || !ForceBottomUp) &&
1465 "-misched-topdown incompatible with -misched-bottomup");
1468 void ConvergingScheduler::releaseTopNode(SUnit *SU) {
1469 if (SU->isScheduled)
1472 for (SUnit::pred_iterator I = SU->Preds.begin(), E = SU->Preds.end();
1474 unsigned PredReadyCycle = I->getSUnit()->TopReadyCycle;
1475 unsigned MinLatency = I->getMinLatency();
1477 Top.MaxMinLatency = std::max(MinLatency, Top.MaxMinLatency);
1479 if (SU->TopReadyCycle < PredReadyCycle + MinLatency)
1480 SU->TopReadyCycle = PredReadyCycle + MinLatency;
1482 Top.releaseNode(SU, SU->TopReadyCycle);
1485 void ConvergingScheduler::releaseBottomNode(SUnit *SU) {
1486 if (SU->isScheduled)
1489 assert(SU->getInstr() && "Scheduled SUnit must have instr");
1491 for (SUnit::succ_iterator I = SU->Succs.begin(), E = SU->Succs.end();
1495 unsigned SuccReadyCycle = I->getSUnit()->BotReadyCycle;
1496 unsigned MinLatency = I->getMinLatency();
1498 Bot.MaxMinLatency = std::max(MinLatency, Bot.MaxMinLatency);
1500 if (SU->BotReadyCycle < SuccReadyCycle + MinLatency)
1501 SU->BotReadyCycle = SuccReadyCycle + MinLatency;
1503 Bot.releaseNode(SU, SU->BotReadyCycle);
1506 void ConvergingScheduler::registerRoots() {
1507 Rem.CriticalPath = DAG->ExitSU.getDepth();
1508 // Some roots may not feed into ExitSU. Check all of them in case.
1509 for (std::vector<SUnit*>::const_iterator
1510 I = Bot.Available.begin(), E = Bot.Available.end(); I != E; ++I) {
1511 if ((*I)->getDepth() > Rem.CriticalPath)
1512 Rem.CriticalPath = (*I)->getDepth();
1514 DEBUG(dbgs() << "Critical Path: " << Rem.CriticalPath << '\n');
1517 /// Does this SU have a hazard within the current instruction group.
1519 /// The scheduler supports two modes of hazard recognition. The first is the
1520 /// ScheduleHazardRecognizer API. It is a fully general hazard recognizer that
1521 /// supports highly complicated in-order reservation tables
1522 /// (ScoreboardHazardRecognizer) and arbitraty target-specific logic.
1524 /// The second is a streamlined mechanism that checks for hazards based on
1525 /// simple counters that the scheduler itself maintains. It explicitly checks
1526 /// for instruction dispatch limitations, including the number of micro-ops that
1527 /// can dispatch per cycle.
1529 /// TODO: Also check whether the SU must start a new group.
1530 bool ConvergingScheduler::SchedBoundary::checkHazard(SUnit *SU) {
1531 if (HazardRec->isEnabled())
1532 return HazardRec->getHazardType(SU) != ScheduleHazardRecognizer::NoHazard;
1534 unsigned uops = SchedModel->getNumMicroOps(SU->getInstr());
1535 if ((IssueCount > 0) && (IssueCount + uops > SchedModel->getIssueWidth())) {
1536 DEBUG(dbgs() << " SU(" << SU->NodeNum << ") uops="
1537 << SchedModel->getNumMicroOps(SU->getInstr()) << '\n');
1543 /// Compute the remaining latency to determine whether ILP should be increased.
1544 void ConvergingScheduler::SchedBoundary::setLatencyPolicy(CandPolicy &Policy) {
1545 DEBUG(dbgs() << " " << Available.getName()
1546 << " DependentLatency " << DependentLatency << '\n');
1548 // FIXME: compile time. In all, we visit four queues here one we should only
1549 // need to visit the one that was last popped if we cache the result.
1550 unsigned RemLatency = DependentLatency;
1551 for (ReadyQueue::iterator I = Available.begin(), E = Available.end();
1553 unsigned L = getUnscheduledLatency(*I);
1554 if (L > RemLatency) {
1555 DEBUG(dbgs() << " " << Available.getName()
1556 << " RemLatency SU(" << (*I)->NodeNum << ") " << L << '\n');
1560 for (ReadyQueue::iterator I = Pending.begin(), E = Pending.end();
1562 unsigned L = getUnscheduledLatency(*I);
1566 unsigned CriticalPathLimit = Rem->CriticalPath + SchedModel->getILPWindow();
1567 DEBUG(dbgs() << " " << Available.getName()
1568 << " ExpectedLatency " << ExpectedLatency
1569 << " CP Limit " << CriticalPathLimit << '\n');
1571 if (RemLatency + std::max(ExpectedLatency, CurrCycle) >= CriticalPathLimit
1572 && RemLatency > Rem->getMaxRemainingCount(SchedModel)) {
1573 Policy.ReduceLatency = true;
1574 DEBUG(dbgs() << " Increase ILP: " << Available.getName() << '\n');
1578 void ConvergingScheduler::SchedBoundary::releaseNode(SUnit *SU,
1579 unsigned ReadyCycle) {
1581 if (ReadyCycle < MinReadyCycle)
1582 MinReadyCycle = ReadyCycle;
1584 // Check for interlocks first. For the purpose of other heuristics, an
1585 // instruction that cannot issue appears as if it's not in the ReadyQueue.
1586 if (ReadyCycle > CurrCycle || checkHazard(SU))
1591 // Record this node as an immediate dependent of the scheduled node.
1595 /// Move the boundary of scheduled code by one cycle.
1596 void ConvergingScheduler::SchedBoundary::bumpCycle() {
1597 unsigned Width = SchedModel->getIssueWidth();
1598 IssueCount = (IssueCount <= Width) ? 0 : IssueCount - Width;
1600 unsigned NextCycle = CurrCycle + 1;
1601 assert(MinReadyCycle < UINT_MAX && "MinReadyCycle uninitialized");
1602 if (MinReadyCycle > NextCycle) {
1604 NextCycle = MinReadyCycle;
1606 if ((NextCycle - CurrCycle) > DependentLatency)
1607 DependentLatency = 0;
1609 DependentLatency -= (NextCycle - CurrCycle);
1611 if (!HazardRec->isEnabled()) {
1612 // Bypass HazardRec virtual calls.
1613 CurrCycle = NextCycle;
1616 // Bypass getHazardType calls in case of long latency.
1617 for (; CurrCycle != NextCycle; ++CurrCycle) {
1619 HazardRec->AdvanceCycle();
1621 HazardRec->RecedeCycle();
1624 CheckPending = true;
1625 IsResourceLimited = getCriticalCount() > std::max(ExpectedLatency, CurrCycle);
1627 DEBUG(dbgs() << " " << Available.getName()
1628 << " Cycle: " << CurrCycle << '\n');
1631 /// Add the given processor resource to this scheduled zone.
1632 void ConvergingScheduler::SchedBoundary::countResource(unsigned PIdx,
1634 unsigned Factor = SchedModel->getResourceFactor(PIdx);
1635 DEBUG(dbgs() << " " << SchedModel->getProcResource(PIdx)->Name
1636 << " +(" << Cycles << "x" << Factor
1637 << ") / " << SchedModel->getLatencyFactor() << '\n');
1639 unsigned Count = Factor * Cycles;
1640 ResourceCounts[PIdx] += Count;
1641 assert(Rem->RemainingCounts[PIdx] >= Count && "resource double counted");
1642 Rem->RemainingCounts[PIdx] -= Count;
1644 // Check if this resource exceeds the current critical resource by a full
1645 // cycle. If so, it becomes the critical resource.
1646 if ((int)(ResourceCounts[PIdx] - ResourceCounts[CritResIdx])
1647 >= (int)SchedModel->getLatencyFactor()) {
1649 DEBUG(dbgs() << " *** Critical resource "
1650 << SchedModel->getProcResource(PIdx)->Name << " x"
1651 << ResourceCounts[PIdx] << '\n');
1655 /// Move the boundary of scheduled code by one SUnit.
1656 void ConvergingScheduler::SchedBoundary::bumpNode(SUnit *SU) {
1657 // Update the reservation table.
1658 if (HazardRec->isEnabled()) {
1659 if (!isTop() && SU->isCall) {
1660 // Calls are scheduled with their preceding instructions. For bottom-up
1661 // scheduling, clear the pipeline state before emitting.
1664 HazardRec->EmitInstruction(SU);
1666 // Update resource counts and critical resource.
1667 if (SchedModel->hasInstrSchedModel()) {
1668 const MCSchedClassDesc *SC = DAG->getSchedClass(SU);
1669 Rem->RemainingMicroOps -= SchedModel->getNumMicroOps(SU->getInstr(), SC);
1670 for (TargetSchedModel::ProcResIter
1671 PI = SchedModel->getWriteProcResBegin(SC),
1672 PE = SchedModel->getWriteProcResEnd(SC); PI != PE; ++PI) {
1673 countResource(PI->ProcResourceIdx, PI->Cycles);
1676 unsigned &TopLatency = isTop() ? ExpectedLatency : DependentLatency;
1677 unsigned &BotLatency = isTop() ? DependentLatency : ExpectedLatency;
1678 if (SU->getDepth() > TopLatency)
1679 TopLatency = SU->getDepth();
1680 if (SU->getHeight() > BotLatency)
1681 BotLatency = SU->getHeight();
1683 IsResourceLimited = getCriticalCount() > std::max(ExpectedLatency, CurrCycle);
1685 // Check the instruction group dispatch limit.
1686 // TODO: Check if this SU must end a dispatch group.
1687 IssueCount += SchedModel->getNumMicroOps(SU->getInstr());
1689 // checkHazard prevents scheduling multiple instructions per cycle that exceed
1690 // issue width. However, we commonly reach the maximum. In this case
1691 // opportunistically bump the cycle to avoid uselessly checking everything in
1692 // the readyQ. Furthermore, a single instruction may produce more than one
1693 // cycle's worth of micro-ops.
1694 if (IssueCount >= SchedModel->getIssueWidth()) {
1695 DEBUG(dbgs() << " *** Max instrs at cycle " << CurrCycle << '\n');
1700 /// Release pending ready nodes in to the available queue. This makes them
1701 /// visible to heuristics.
1702 void ConvergingScheduler::SchedBoundary::releasePending() {
1703 // If the available queue is empty, it is safe to reset MinReadyCycle.
1704 if (Available.empty())
1705 MinReadyCycle = UINT_MAX;
1707 // Check to see if any of the pending instructions are ready to issue. If
1708 // so, add them to the available queue.
1709 for (unsigned i = 0, e = Pending.size(); i != e; ++i) {
1710 SUnit *SU = *(Pending.begin()+i);
1711 unsigned ReadyCycle = isTop() ? SU->TopReadyCycle : SU->BotReadyCycle;
1713 if (ReadyCycle < MinReadyCycle)
1714 MinReadyCycle = ReadyCycle;
1716 if (ReadyCycle > CurrCycle)
1719 if (checkHazard(SU))
1723 Pending.remove(Pending.begin()+i);
1726 DEBUG(if (!Pending.empty()) Pending.dump());
1727 CheckPending = false;
1730 /// Remove SU from the ready set for this boundary.
1731 void ConvergingScheduler::SchedBoundary::removeReady(SUnit *SU) {
1732 if (Available.isInQueue(SU))
1733 Available.remove(Available.find(SU));
1735 assert(Pending.isInQueue(SU) && "bad ready count");
1736 Pending.remove(Pending.find(SU));
1740 /// If this queue only has one ready candidate, return it. As a side effect,
1741 /// defer any nodes that now hit a hazard, and advance the cycle until at least
1742 /// one node is ready. If multiple instructions are ready, return NULL.
1743 SUnit *ConvergingScheduler::SchedBoundary::pickOnlyChoice() {
1747 if (IssueCount > 0) {
1748 // Defer any ready instrs that now have a hazard.
1749 for (ReadyQueue::iterator I = Available.begin(); I != Available.end();) {
1750 if (checkHazard(*I)) {
1752 I = Available.remove(I);
1758 for (unsigned i = 0; Available.empty(); ++i) {
1759 assert(i <= (HazardRec->getMaxLookAhead() + MaxMinLatency) &&
1760 "permanent hazard"); (void)i;
1764 if (Available.size() == 1)
1765 return *Available.begin();
1769 /// Record the candidate policy for opposite zones with different critical
1772 /// If the CriticalZone is latency limited, don't force a policy for the
1773 /// candidates here. Instead, setLatencyPolicy sets ReduceLatency if needed.
1774 void ConvergingScheduler::balanceZones(
1775 ConvergingScheduler::SchedBoundary &CriticalZone,
1776 ConvergingScheduler::SchedCandidate &CriticalCand,
1777 ConvergingScheduler::SchedBoundary &OppositeZone,
1778 ConvergingScheduler::SchedCandidate &OppositeCand) {
1780 if (!CriticalZone.IsResourceLimited)
1782 assert(SchedModel->hasInstrSchedModel() && "required schedmodel");
1784 SchedRemainder *Rem = CriticalZone.Rem;
1786 // If the critical zone is overconsuming a resource relative to the
1787 // remainder, try to reduce it.
1788 unsigned RemainingCritCount =
1789 Rem->RemainingCounts[CriticalZone.CritResIdx];
1790 if ((int)(Rem->getMaxRemainingCount(SchedModel) - RemainingCritCount)
1791 > (int)SchedModel->getLatencyFactor()) {
1792 CriticalCand.Policy.ReduceResIdx = CriticalZone.CritResIdx;
1793 DEBUG(dbgs() << " Balance " << CriticalZone.Available.getName()
1795 << SchedModel->getProcResource(CriticalZone.CritResIdx)->Name
1798 // If the other zone is underconsuming a resource relative to the full zone,
1799 // try to increase it.
1800 unsigned OppositeCount =
1801 OppositeZone.ResourceCounts[CriticalZone.CritResIdx];
1802 if ((int)(OppositeZone.ExpectedCount - OppositeCount)
1803 > (int)SchedModel->getLatencyFactor()) {
1804 OppositeCand.Policy.DemandResIdx = CriticalZone.CritResIdx;
1805 DEBUG(dbgs() << " Balance " << OppositeZone.Available.getName()
1807 << SchedModel->getProcResource(OppositeZone.CritResIdx)->Name
1812 /// Determine if the scheduled zones exceed resource limits or critical path and
1813 /// set each candidate's ReduceHeight policy accordingly.
1814 void ConvergingScheduler::checkResourceLimits(
1815 ConvergingScheduler::SchedCandidate &TopCand,
1816 ConvergingScheduler::SchedCandidate &BotCand) {
1818 // Set ReduceLatency to true if needed.
1819 Bot.setLatencyPolicy(BotCand.Policy);
1820 Top.setLatencyPolicy(TopCand.Policy);
1822 // Handle resource-limited regions.
1823 if (Top.IsResourceLimited && Bot.IsResourceLimited
1824 && Top.CritResIdx == Bot.CritResIdx) {
1825 // If the scheduled critical resource in both zones is no longer the
1826 // critical remaining resource, attempt to reduce resource height both ways.
1827 if (Top.CritResIdx != Rem.CritResIdx) {
1828 TopCand.Policy.ReduceResIdx = Top.CritResIdx;
1829 BotCand.Policy.ReduceResIdx = Bot.CritResIdx;
1830 DEBUG(dbgs() << " Reduce scheduled "
1831 << SchedModel->getProcResource(Top.CritResIdx)->Name << '\n');
1835 // Handle latency-limited regions.
1836 if (!Top.IsResourceLimited && !Bot.IsResourceLimited) {
1837 // If the total scheduled expected latency exceeds the region's critical
1838 // path then reduce latency both ways.
1840 // Just because a zone is not resource limited does not mean it is latency
1841 // limited. Unbuffered resource, such as max micro-ops may cause CurrCycle
1842 // to exceed expected latency.
1843 if ((Top.ExpectedLatency + Bot.ExpectedLatency >= Rem.CriticalPath)
1844 && (Rem.CriticalPath > Top.CurrCycle + Bot.CurrCycle)) {
1845 TopCand.Policy.ReduceLatency = true;
1846 BotCand.Policy.ReduceLatency = true;
1847 DEBUG(dbgs() << " Reduce scheduled latency " << Top.ExpectedLatency
1848 << " + " << Bot.ExpectedLatency << '\n');
1852 // The critical resource is different in each zone, so request balancing.
1854 // Compute the cost of each zone.
1855 Top.ExpectedCount = std::max(Top.ExpectedLatency, Top.CurrCycle);
1856 Top.ExpectedCount = std::max(
1857 Top.getCriticalCount(),
1858 Top.ExpectedCount * SchedModel->getLatencyFactor());
1859 Bot.ExpectedCount = std::max(Bot.ExpectedLatency, Bot.CurrCycle);
1860 Bot.ExpectedCount = std::max(
1861 Bot.getCriticalCount(),
1862 Bot.ExpectedCount * SchedModel->getLatencyFactor());
1864 balanceZones(Top, TopCand, Bot, BotCand);
1865 balanceZones(Bot, BotCand, Top, TopCand);
1868 void ConvergingScheduler::SchedCandidate::
1869 initResourceDelta(const ScheduleDAGMI *DAG,
1870 const TargetSchedModel *SchedModel) {
1871 if (!Policy.ReduceResIdx && !Policy.DemandResIdx)
1874 const MCSchedClassDesc *SC = DAG->getSchedClass(SU);
1875 for (TargetSchedModel::ProcResIter
1876 PI = SchedModel->getWriteProcResBegin(SC),
1877 PE = SchedModel->getWriteProcResEnd(SC); PI != PE; ++PI) {
1878 if (PI->ProcResourceIdx == Policy.ReduceResIdx)
1879 ResDelta.CritResources += PI->Cycles;
1880 if (PI->ProcResourceIdx == Policy.DemandResIdx)
1881 ResDelta.DemandedResources += PI->Cycles;
1885 /// Return true if this heuristic determines order.
1886 static bool tryLess(int TryVal, int CandVal,
1887 ConvergingScheduler::SchedCandidate &TryCand,
1888 ConvergingScheduler::SchedCandidate &Cand,
1889 ConvergingScheduler::CandReason Reason) {
1890 if (TryVal < CandVal) {
1891 TryCand.Reason = Reason;
1894 if (TryVal > CandVal) {
1895 if (Cand.Reason > Reason)
1896 Cand.Reason = Reason;
1902 static bool tryGreater(int TryVal, int CandVal,
1903 ConvergingScheduler::SchedCandidate &TryCand,
1904 ConvergingScheduler::SchedCandidate &Cand,
1905 ConvergingScheduler::CandReason Reason) {
1906 if (TryVal > CandVal) {
1907 TryCand.Reason = Reason;
1910 if (TryVal < CandVal) {
1911 if (Cand.Reason > Reason)
1912 Cand.Reason = Reason;
1918 static unsigned getWeakLeft(const SUnit *SU, bool isTop) {
1919 return (isTop) ? SU->WeakPredsLeft : SU->WeakSuccsLeft;
1922 /// Minimize physical register live ranges. Regalloc wants them adjacent to
1923 /// their physreg def/use.
1925 /// FIXME: This is an unnecessary check on the critical path. Most are root/leaf
1926 /// copies which can be prescheduled. The rest (e.g. x86 MUL) could be bundled
1927 /// with the operation that produces or consumes the physreg. We'll do this when
1928 /// regalloc has support for parallel copies.
1929 static int biasPhysRegCopy(const SUnit *SU, bool isTop) {
1930 const MachineInstr *MI = SU->getInstr();
1934 unsigned ScheduledOper = isTop ? 1 : 0;
1935 unsigned UnscheduledOper = isTop ? 0 : 1;
1936 // If we have already scheduled the physreg produce/consumer, immediately
1937 // schedule the copy.
1938 if (TargetRegisterInfo::isPhysicalRegister(
1939 MI->getOperand(ScheduledOper).getReg()))
1941 // If the physreg is at the boundary, defer it. Otherwise schedule it
1942 // immediately to free the dependent. We can hoist the copy later.
1943 bool AtBoundary = isTop ? !SU->NumSuccsLeft : !SU->NumPredsLeft;
1944 if (TargetRegisterInfo::isPhysicalRegister(
1945 MI->getOperand(UnscheduledOper).getReg()))
1946 return AtBoundary ? -1 : 1;
1950 /// Apply a set of heursitics to a new candidate. Heuristics are currently
1951 /// hierarchical. This may be more efficient than a graduated cost model because
1952 /// we don't need to evaluate all aspects of the model for each node in the
1953 /// queue. But it's really done to make the heuristics easier to debug and
1954 /// statistically analyze.
1956 /// \param Cand provides the policy and current best candidate.
1957 /// \param TryCand refers to the next SUnit candidate, otherwise uninitialized.
1958 /// \param Zone describes the scheduled zone that we are extending.
1959 /// \param RPTracker describes reg pressure within the scheduled zone.
1960 /// \param TempTracker is a scratch pressure tracker to reuse in queries.
1961 void ConvergingScheduler::tryCandidate(SchedCandidate &Cand,
1962 SchedCandidate &TryCand,
1963 SchedBoundary &Zone,
1964 const RegPressureTracker &RPTracker,
1965 RegPressureTracker &TempTracker) {
1967 // Always initialize TryCand's RPDelta.
1968 TempTracker.getMaxPressureDelta(TryCand.SU->getInstr(), TryCand.RPDelta,
1969 DAG->getRegionCriticalPSets(),
1970 DAG->getRegPressure().MaxSetPressure);
1972 // Initialize the candidate if needed.
1973 if (!Cand.isValid()) {
1974 TryCand.Reason = NodeOrder;
1978 if (tryGreater(biasPhysRegCopy(TryCand.SU, Zone.isTop()),
1979 biasPhysRegCopy(Cand.SU, Zone.isTop()),
1980 TryCand, Cand, PhysRegCopy))
1983 // Avoid exceeding the target's limit.
1984 if (tryLess(TryCand.RPDelta.Excess.UnitIncrease,
1985 Cand.RPDelta.Excess.UnitIncrease, TryCand, Cand, SingleExcess))
1987 if (Cand.Reason == SingleExcess)
1988 Cand.Reason = MultiPressure;
1990 // Avoid increasing the max critical pressure in the scheduled region.
1991 if (tryLess(TryCand.RPDelta.CriticalMax.UnitIncrease,
1992 Cand.RPDelta.CriticalMax.UnitIncrease,
1993 TryCand, Cand, SingleCritical))
1995 if (Cand.Reason == SingleCritical)
1996 Cand.Reason = MultiPressure;
1998 // Keep clustered nodes together to encourage downstream peephole
1999 // optimizations which may reduce resource requirements.
2001 // This is a best effort to set things up for a post-RA pass. Optimizations
2002 // like generating loads of multiple registers should ideally be done within
2003 // the scheduler pass by combining the loads during DAG postprocessing.
2004 const SUnit *NextClusterSU =
2005 Zone.isTop() ? DAG->getNextClusterSucc() : DAG->getNextClusterPred();
2006 if (tryGreater(TryCand.SU == NextClusterSU, Cand.SU == NextClusterSU,
2007 TryCand, Cand, Cluster))
2010 // Weak edges are for clustering and other constraints.
2012 // Deferring TryCand here does not change Cand's reason. This is good in the
2013 // sense that a bad candidate shouldn't affect a previous candidate's
2014 // goodness, but bad in that it is assymetric and depends on queue order.
2015 CandReason OrigReason = Cand.Reason;
2016 if (tryLess(getWeakLeft(TryCand.SU, Zone.isTop()),
2017 getWeakLeft(Cand.SU, Zone.isTop()),
2018 TryCand, Cand, Weak)) {
2019 Cand.Reason = OrigReason;
2022 // Avoid critical resource consumption and balance the schedule.
2023 TryCand.initResourceDelta(DAG, SchedModel);
2024 if (tryLess(TryCand.ResDelta.CritResources, Cand.ResDelta.CritResources,
2025 TryCand, Cand, ResourceReduce))
2027 if (tryGreater(TryCand.ResDelta.DemandedResources,
2028 Cand.ResDelta.DemandedResources,
2029 TryCand, Cand, ResourceDemand))
2032 // Avoid serializing long latency dependence chains.
2033 if (Cand.Policy.ReduceLatency) {
2035 if (Cand.SU->getDepth() * SchedModel->getLatencyFactor()
2036 > Zone.ExpectedCount) {
2037 if (tryLess(TryCand.SU->getDepth(), Cand.SU->getDepth(),
2038 TryCand, Cand, TopDepthReduce))
2041 if (tryGreater(TryCand.SU->getHeight(), Cand.SU->getHeight(),
2042 TryCand, Cand, TopPathReduce))
2046 if (Cand.SU->getHeight() * SchedModel->getLatencyFactor()
2047 > Zone.ExpectedCount) {
2048 if (tryLess(TryCand.SU->getHeight(), Cand.SU->getHeight(),
2049 TryCand, Cand, BotHeightReduce))
2052 if (tryGreater(TryCand.SU->getDepth(), Cand.SU->getDepth(),
2053 TryCand, Cand, BotPathReduce))
2058 // Avoid increasing the max pressure of the entire region.
2059 if (tryLess(TryCand.RPDelta.CurrentMax.UnitIncrease,
2060 Cand.RPDelta.CurrentMax.UnitIncrease, TryCand, Cand, SingleMax))
2062 if (Cand.Reason == SingleMax)
2063 Cand.Reason = MultiPressure;
2065 // Prefer immediate defs/users of the last scheduled instruction. This is a
2066 // nice pressure avoidance strategy that also conserves the processor's
2067 // register renaming resources and keeps the machine code readable.
2068 if (tryGreater(Zone.NextSUs.count(TryCand.SU), Zone.NextSUs.count(Cand.SU),
2069 TryCand, Cand, NextDefUse))
2072 // Fall through to original instruction order.
2073 if ((Zone.isTop() && TryCand.SU->NodeNum < Cand.SU->NodeNum)
2074 || (!Zone.isTop() && TryCand.SU->NodeNum > Cand.SU->NodeNum)) {
2075 TryCand.Reason = NodeOrder;
2079 /// pickNodeFromQueue helper that returns true if the LHS reg pressure effect is
2080 /// more desirable than RHS from scheduling standpoint.
2081 static bool compareRPDelta(const RegPressureDelta &LHS,
2082 const RegPressureDelta &RHS) {
2083 // Compare each component of pressure in decreasing order of importance
2084 // without checking if any are valid. Invalid PressureElements are assumed to
2085 // have UnitIncrease==0, so are neutral.
2087 // Avoid increasing the max critical pressure in the scheduled region.
2088 if (LHS.Excess.UnitIncrease != RHS.Excess.UnitIncrease) {
2089 DEBUG(dbgs() << " RP excess top - bot: "
2090 << (LHS.Excess.UnitIncrease - RHS.Excess.UnitIncrease) << '\n');
2091 return LHS.Excess.UnitIncrease < RHS.Excess.UnitIncrease;
2093 // Avoid increasing the max critical pressure in the scheduled region.
2094 if (LHS.CriticalMax.UnitIncrease != RHS.CriticalMax.UnitIncrease) {
2095 DEBUG(dbgs() << " RP critical top - bot: "
2096 << (LHS.CriticalMax.UnitIncrease - RHS.CriticalMax.UnitIncrease)
2098 return LHS.CriticalMax.UnitIncrease < RHS.CriticalMax.UnitIncrease;
2100 // Avoid increasing the max pressure of the entire region.
2101 if (LHS.CurrentMax.UnitIncrease != RHS.CurrentMax.UnitIncrease) {
2102 DEBUG(dbgs() << " RP current top - bot: "
2103 << (LHS.CurrentMax.UnitIncrease - RHS.CurrentMax.UnitIncrease)
2105 return LHS.CurrentMax.UnitIncrease < RHS.CurrentMax.UnitIncrease;
2111 const char *ConvergingScheduler::getReasonStr(
2112 ConvergingScheduler::CandReason Reason) {
2114 case NoCand: return "NOCAND ";
2115 case PhysRegCopy: return "PREG-COPY";
2116 case SingleExcess: return "REG-EXCESS";
2117 case SingleCritical: return "REG-CRIT ";
2118 case Cluster: return "CLUSTER ";
2119 case Weak: return "WEAK ";
2120 case SingleMax: return "REG-MAX ";
2121 case MultiPressure: return "REG-MULTI ";
2122 case ResourceReduce: return "RES-REDUCE";
2123 case ResourceDemand: return "RES-DEMAND";
2124 case TopDepthReduce: return "TOP-DEPTH ";
2125 case TopPathReduce: return "TOP-PATH ";
2126 case BotHeightReduce:return "BOT-HEIGHT";
2127 case BotPathReduce: return "BOT-PATH ";
2128 case NextDefUse: return "DEF-USE ";
2129 case NodeOrder: return "ORDER ";
2131 llvm_unreachable("Unknown reason!");
2134 void ConvergingScheduler::traceCandidate(const SchedCandidate &Cand) {
2136 unsigned ResIdx = 0;
2137 unsigned Latency = 0;
2138 switch (Cand.Reason) {
2142 P = Cand.RPDelta.Excess;
2144 case SingleCritical:
2145 P = Cand.RPDelta.CriticalMax;
2148 P = Cand.RPDelta.CurrentMax;
2150 case ResourceReduce:
2151 ResIdx = Cand.Policy.ReduceResIdx;
2153 case ResourceDemand:
2154 ResIdx = Cand.Policy.DemandResIdx;
2156 case TopDepthReduce:
2157 Latency = Cand.SU->getDepth();
2160 Latency = Cand.SU->getHeight();
2162 case BotHeightReduce:
2163 Latency = Cand.SU->getHeight();
2166 Latency = Cand.SU->getDepth();
2169 dbgs() << " SU(" << Cand.SU->NodeNum << ") " << getReasonStr(Cand.Reason);
2171 dbgs() << " " << TRI->getRegPressureSetName(P.PSetID)
2172 << ":" << P.UnitIncrease << " ";
2176 dbgs() << " " << SchedModel->getProcResource(ResIdx)->Name << " ";
2180 dbgs() << " " << Latency << " cycles ";
2187 /// Pick the best candidate from the top queue.
2189 /// TODO: getMaxPressureDelta results can be mostly cached for each SUnit during
2190 /// DAG building. To adjust for the current scheduling location we need to
2191 /// maintain the number of vreg uses remaining to be top-scheduled.
2192 void ConvergingScheduler::pickNodeFromQueue(SchedBoundary &Zone,
2193 const RegPressureTracker &RPTracker,
2194 SchedCandidate &Cand) {
2195 ReadyQueue &Q = Zone.Available;
2199 // getMaxPressureDelta temporarily modifies the tracker.
2200 RegPressureTracker &TempTracker = const_cast<RegPressureTracker&>(RPTracker);
2202 for (ReadyQueue::iterator I = Q.begin(), E = Q.end(); I != E; ++I) {
2204 SchedCandidate TryCand(Cand.Policy);
2206 tryCandidate(Cand, TryCand, Zone, RPTracker, TempTracker);
2207 if (TryCand.Reason != NoCand) {
2208 // Initialize resource delta if needed in case future heuristics query it.
2209 if (TryCand.ResDelta == SchedResourceDelta())
2210 TryCand.initResourceDelta(DAG, SchedModel);
2211 Cand.setBest(TryCand);
2212 DEBUG(traceCandidate(Cand));
2217 static void tracePick(const ConvergingScheduler::SchedCandidate &Cand,
2219 DEBUG(dbgs() << "Pick " << (IsTop ? "Top " : "Bot ")
2220 << ConvergingScheduler::getReasonStr(Cand.Reason) << '\n');
2223 /// Pick the best candidate node from either the top or bottom queue.
2224 SUnit *ConvergingScheduler::pickNodeBidirectional(bool &IsTopNode) {
2225 // Schedule as far as possible in the direction of no choice. This is most
2226 // efficient, but also provides the best heuristics for CriticalPSets.
2227 if (SUnit *SU = Bot.pickOnlyChoice()) {
2229 DEBUG(dbgs() << "Pick Top NOCAND\n");
2232 if (SUnit *SU = Top.pickOnlyChoice()) {
2234 DEBUG(dbgs() << "Pick Bot NOCAND\n");
2237 CandPolicy NoPolicy;
2238 SchedCandidate BotCand(NoPolicy);
2239 SchedCandidate TopCand(NoPolicy);
2240 checkResourceLimits(TopCand, BotCand);
2242 // Prefer bottom scheduling when heuristics are silent.
2243 pickNodeFromQueue(Bot, DAG->getBotRPTracker(), BotCand);
2244 assert(BotCand.Reason != NoCand && "failed to find the first candidate");
2246 // If either Q has a single candidate that provides the least increase in
2247 // Excess pressure, we can immediately schedule from that Q.
2249 // RegionCriticalPSets summarizes the pressure within the scheduled region and
2250 // affects picking from either Q. If scheduling in one direction must
2251 // increase pressure for one of the excess PSets, then schedule in that
2252 // direction first to provide more freedom in the other direction.
2253 if (BotCand.Reason == SingleExcess || BotCand.Reason == SingleCritical) {
2255 tracePick(BotCand, IsTopNode);
2258 // Check if the top Q has a better candidate.
2259 pickNodeFromQueue(Top, DAG->getTopRPTracker(), TopCand);
2260 assert(TopCand.Reason != NoCand && "failed to find the first candidate");
2262 // If either Q has a single candidate that minimizes pressure above the
2263 // original region's pressure pick it.
2264 if (TopCand.Reason <= SingleMax || BotCand.Reason <= SingleMax) {
2265 if (TopCand.Reason < BotCand.Reason) {
2267 tracePick(TopCand, IsTopNode);
2271 tracePick(BotCand, IsTopNode);
2274 // Check for a salient pressure difference and pick the best from either side.
2275 if (compareRPDelta(TopCand.RPDelta, BotCand.RPDelta)) {
2277 tracePick(TopCand, IsTopNode);
2280 // Otherwise prefer the bottom candidate, in node order if all else failed.
2281 if (TopCand.Reason < BotCand.Reason) {
2283 tracePick(TopCand, IsTopNode);
2287 tracePick(BotCand, IsTopNode);
2291 /// Pick the best node to balance the schedule. Implements MachineSchedStrategy.
2292 SUnit *ConvergingScheduler::pickNode(bool &IsTopNode) {
2293 if (DAG->top() == DAG->bottom()) {
2294 assert(Top.Available.empty() && Top.Pending.empty() &&
2295 Bot.Available.empty() && Bot.Pending.empty() && "ReadyQ garbage");
2301 SU = Top.pickOnlyChoice();
2303 CandPolicy NoPolicy;
2304 SchedCandidate TopCand(NoPolicy);
2305 pickNodeFromQueue(Top, DAG->getTopRPTracker(), TopCand);
2306 assert(TopCand.Reason != NoCand && "failed to find the first candidate");
2311 else if (ForceBottomUp) {
2312 SU = Bot.pickOnlyChoice();
2314 CandPolicy NoPolicy;
2315 SchedCandidate BotCand(NoPolicy);
2316 pickNodeFromQueue(Bot, DAG->getBotRPTracker(), BotCand);
2317 assert(BotCand.Reason != NoCand && "failed to find the first candidate");
2323 SU = pickNodeBidirectional(IsTopNode);
2325 } while (SU->isScheduled);
2327 if (SU->isTopReady())
2328 Top.removeReady(SU);
2329 if (SU->isBottomReady())
2330 Bot.removeReady(SU);
2332 DEBUG(dbgs() << "Scheduling SU(" << SU->NodeNum << ") " << *SU->getInstr());
2336 void ConvergingScheduler::reschedulePhysRegCopies(SUnit *SU, bool isTop) {
2338 MachineBasicBlock::iterator InsertPos = SU->getInstr();
2341 SmallVectorImpl<SDep> &Deps = isTop ? SU->Preds : SU->Succs;
2343 // Find already scheduled copies with a single physreg dependence and move
2344 // them just above the scheduled instruction.
2345 for (SmallVectorImpl<SDep>::iterator I = Deps.begin(), E = Deps.end();
2347 if (I->getKind() != SDep::Data || !TRI->isPhysicalRegister(I->getReg()))
2349 SUnit *DepSU = I->getSUnit();
2350 if (isTop ? DepSU->Succs.size() > 1 : DepSU->Preds.size() > 1)
2352 MachineInstr *Copy = DepSU->getInstr();
2353 if (!Copy->isCopy())
2355 DEBUG(dbgs() << " Rescheduling physreg copy ";
2356 I->getSUnit()->dump(DAG));
2357 DAG->moveInstruction(Copy, InsertPos);
2361 /// Update the scheduler's state after scheduling a node. This is the same node
2362 /// that was just returned by pickNode(). However, ScheduleDAGMI needs to update
2363 /// it's state based on the current cycle before MachineSchedStrategy does.
2365 /// FIXME: Eventually, we may bundle physreg copies rather than rescheduling
2366 /// them here. See comments in biasPhysRegCopy.
2367 void ConvergingScheduler::schedNode(SUnit *SU, bool IsTopNode) {
2369 SU->TopReadyCycle = Top.CurrCycle;
2371 if (SU->hasPhysRegUses)
2372 reschedulePhysRegCopies(SU, true);
2375 SU->BotReadyCycle = Bot.CurrCycle;
2377 if (SU->hasPhysRegDefs)
2378 reschedulePhysRegCopies(SU, false);
2382 /// Create the standard converging machine scheduler. This will be used as the
2383 /// default scheduler if the target does not set a default.
2384 static ScheduleDAGInstrs *createConvergingSched(MachineSchedContext *C) {
2385 assert((!ForceTopDown || !ForceBottomUp) &&
2386 "-misched-topdown incompatible with -misched-bottomup");
2387 ScheduleDAGMI *DAG = new ScheduleDAGMI(C, new ConvergingScheduler());
2388 // Register DAG post-processors.
2390 // FIXME: extend the mutation API to allow earlier mutations to instantiate
2391 // data and pass it to later mutations. Have a single mutation that gathers
2392 // the interesting nodes in one pass.
2393 if (EnableCopyConstrain)
2394 DAG->addMutation(new CopyConstrain(DAG->TII, DAG->TRI));
2395 if (EnableLoadCluster)
2396 DAG->addMutation(new LoadClusterMutation(DAG->TII, DAG->TRI));
2397 if (EnableMacroFusion)
2398 DAG->addMutation(new MacroFusion(DAG->TII));
2401 static MachineSchedRegistry
2402 ConvergingSchedRegistry("converge", "Standard converging scheduler.",
2403 createConvergingSched);
2405 //===----------------------------------------------------------------------===//
2406 // ILP Scheduler. Currently for experimental analysis of heuristics.
2407 //===----------------------------------------------------------------------===//
2410 /// \brief Order nodes by the ILP metric.
2412 const SchedDFSResult *DFSResult;
2413 const BitVector *ScheduledTrees;
2416 ILPOrder(bool MaxILP): DFSResult(0), ScheduledTrees(0), MaximizeILP(MaxILP) {}
2418 /// \brief Apply a less-than relation on node priority.
2420 /// (Return true if A comes after B in the Q.)
2421 bool operator()(const SUnit *A, const SUnit *B) const {
2422 unsigned SchedTreeA = DFSResult->getSubtreeID(A);
2423 unsigned SchedTreeB = DFSResult->getSubtreeID(B);
2424 if (SchedTreeA != SchedTreeB) {
2425 // Unscheduled trees have lower priority.
2426 if (ScheduledTrees->test(SchedTreeA) != ScheduledTrees->test(SchedTreeB))
2427 return ScheduledTrees->test(SchedTreeB);
2429 // Trees with shallower connections have have lower priority.
2430 if (DFSResult->getSubtreeLevel(SchedTreeA)
2431 != DFSResult->getSubtreeLevel(SchedTreeB)) {
2432 return DFSResult->getSubtreeLevel(SchedTreeA)
2433 < DFSResult->getSubtreeLevel(SchedTreeB);
2437 return DFSResult->getILP(A) < DFSResult->getILP(B);
2439 return DFSResult->getILP(A) > DFSResult->getILP(B);
2443 /// \brief Schedule based on the ILP metric.
2444 class ILPScheduler : public MachineSchedStrategy {
2445 /// In case all subtrees are eventually connected to a common root through
2446 /// data dependence (e.g. reduction), place an upper limit on their size.
2448 /// FIXME: A subtree limit is generally good, but in the situation commented
2449 /// above, where multiple similar subtrees feed a common root, we should
2450 /// only split at a point where the resulting subtrees will be balanced.
2451 /// (a motivating test case must be found).
2452 static const unsigned SubtreeLimit = 16;
2457 std::vector<SUnit*> ReadyQ;
2459 ILPScheduler(bool MaximizeILP): DAG(0), Cmp(MaximizeILP) {}
2461 virtual void initialize(ScheduleDAGMI *dag) {
2463 DAG->computeDFSResult();
2464 Cmp.DFSResult = DAG->getDFSResult();
2465 Cmp.ScheduledTrees = &DAG->getScheduledTrees();
2469 virtual void registerRoots() {
2470 // Restore the heap in ReadyQ with the updated DFS results.
2471 std::make_heap(ReadyQ.begin(), ReadyQ.end(), Cmp);
2474 /// Implement MachineSchedStrategy interface.
2475 /// -----------------------------------------
2477 /// Callback to select the highest priority node from the ready Q.
2478 virtual SUnit *pickNode(bool &IsTopNode) {
2479 if (ReadyQ.empty()) return NULL;
2480 std::pop_heap(ReadyQ.begin(), ReadyQ.end(), Cmp);
2481 SUnit *SU = ReadyQ.back();
2484 DEBUG(dbgs() << "Pick node " << "SU(" << SU->NodeNum << ") "
2485 << " ILP: " << DAG->getDFSResult()->getILP(SU)
2486 << " Tree: " << DAG->getDFSResult()->getSubtreeID(SU) << " @"
2487 << DAG->getDFSResult()->getSubtreeLevel(
2488 DAG->getDFSResult()->getSubtreeID(SU)) << '\n'
2489 << "Scheduling " << *SU->getInstr());
2493 /// \brief Scheduler callback to notify that a new subtree is scheduled.
2494 virtual void scheduleTree(unsigned SubtreeID) {
2495 std::make_heap(ReadyQ.begin(), ReadyQ.end(), Cmp);
2498 /// Callback after a node is scheduled. Mark a newly scheduled tree, notify
2499 /// DFSResults, and resort the priority Q.
2500 virtual void schedNode(SUnit *SU, bool IsTopNode) {
2501 assert(!IsTopNode && "SchedDFSResult needs bottom-up");
2504 virtual void releaseTopNode(SUnit *) { /*only called for top roots*/ }
2506 virtual void releaseBottomNode(SUnit *SU) {
2507 ReadyQ.push_back(SU);
2508 std::push_heap(ReadyQ.begin(), ReadyQ.end(), Cmp);
2513 static ScheduleDAGInstrs *createILPMaxScheduler(MachineSchedContext *C) {
2514 return new ScheduleDAGMI(C, new ILPScheduler(true));
2516 static ScheduleDAGInstrs *createILPMinScheduler(MachineSchedContext *C) {
2517 return new ScheduleDAGMI(C, new ILPScheduler(false));
2519 static MachineSchedRegistry ILPMaxRegistry(
2520 "ilpmax", "Schedule bottom-up for max ILP", createILPMaxScheduler);
2521 static MachineSchedRegistry ILPMinRegistry(
2522 "ilpmin", "Schedule bottom-up for min ILP", createILPMinScheduler);
2524 //===----------------------------------------------------------------------===//
2525 // Machine Instruction Shuffler for Correctness Testing
2526 //===----------------------------------------------------------------------===//
2530 /// Apply a less-than relation on the node order, which corresponds to the
2531 /// instruction order prior to scheduling. IsReverse implements greater-than.
2532 template<bool IsReverse>
2534 bool operator()(SUnit *A, SUnit *B) const {
2536 return A->NodeNum > B->NodeNum;
2538 return A->NodeNum < B->NodeNum;
2542 /// Reorder instructions as much as possible.
2543 class InstructionShuffler : public MachineSchedStrategy {
2547 // Using a less-than relation (SUnitOrder<false>) for the TopQ priority
2548 // gives nodes with a higher number higher priority causing the latest
2549 // instructions to be scheduled first.
2550 PriorityQueue<SUnit*, std::vector<SUnit*>, SUnitOrder<false> >
2552 // When scheduling bottom-up, use greater-than as the queue priority.
2553 PriorityQueue<SUnit*, std::vector<SUnit*>, SUnitOrder<true> >
2556 InstructionShuffler(bool alternate, bool topdown)
2557 : IsAlternating(alternate), IsTopDown(topdown) {}
2559 virtual void initialize(ScheduleDAGMI *) {
2564 /// Implement MachineSchedStrategy interface.
2565 /// -----------------------------------------
2567 virtual SUnit *pickNode(bool &IsTopNode) {
2571 if (TopQ.empty()) return NULL;
2574 } while (SU->isScheduled);
2579 if (BottomQ.empty()) return NULL;
2582 } while (SU->isScheduled);
2586 IsTopDown = !IsTopDown;
2590 virtual void schedNode(SUnit *SU, bool IsTopNode) {}
2592 virtual void releaseTopNode(SUnit *SU) {
2595 virtual void releaseBottomNode(SUnit *SU) {
2601 static ScheduleDAGInstrs *createInstructionShuffler(MachineSchedContext *C) {
2602 bool Alternate = !ForceTopDown && !ForceBottomUp;
2603 bool TopDown = !ForceBottomUp;
2604 assert((TopDown || !ForceTopDown) &&
2605 "-misched-topdown incompatible with -misched-bottomup");
2606 return new ScheduleDAGMI(C, new InstructionShuffler(Alternate, TopDown));
2608 static MachineSchedRegistry ShufflerRegistry(
2609 "shuffle", "Shuffle machine instructions alternating directions",
2610 createInstructionShuffler);
2613 //===----------------------------------------------------------------------===//
2614 // GraphWriter support for ScheduleDAGMI.
2615 //===----------------------------------------------------------------------===//
2620 template<> struct GraphTraits<
2621 ScheduleDAGMI*> : public GraphTraits<ScheduleDAG*> {};
2624 struct DOTGraphTraits<ScheduleDAGMI*> : public DefaultDOTGraphTraits {
2626 DOTGraphTraits (bool isSimple=false) : DefaultDOTGraphTraits(isSimple) {}
2628 static std::string getGraphName(const ScheduleDAG *G) {
2629 return G->MF.getName();
2632 static bool renderGraphFromBottomUp() {
2636 static bool isNodeHidden(const SUnit *Node) {
2637 return (Node->NumPreds > 10 || Node->NumSuccs > 10);
2640 static bool hasNodeAddressLabel(const SUnit *Node,
2641 const ScheduleDAG *Graph) {
2645 /// If you want to override the dot attributes printed for a particular
2646 /// edge, override this method.
2647 static std::string getEdgeAttributes(const SUnit *Node,
2649 const ScheduleDAG *Graph) {
2650 if (EI.isArtificialDep())
2651 return "color=cyan,style=dashed";
2653 return "color=blue,style=dashed";
2657 static std::string getNodeLabel(const SUnit *SU, const ScheduleDAG *G) {
2659 raw_string_ostream SS(Str);
2660 SS << "SU(" << SU->NodeNum << ')';
2663 static std::string getNodeDescription(const SUnit *SU, const ScheduleDAG *G) {
2664 return G->getGraphNodeLabel(SU);
2667 static std::string getNodeAttributes(const SUnit *N,
2668 const ScheduleDAG *Graph) {
2669 std::string Str("shape=Mrecord");
2670 const SchedDFSResult *DFS =
2671 static_cast<const ScheduleDAGMI*>(Graph)->getDFSResult();
2673 Str += ",style=filled,fillcolor=\"#";
2674 Str += DOT::getColorString(DFS->getSubtreeID(N));
2683 /// viewGraph - Pop up a ghostview window with the reachable parts of the DAG
2684 /// rendered using 'dot'.
2686 void ScheduleDAGMI::viewGraph(const Twine &Name, const Twine &Title) {
2688 ViewGraph(this, Name, false, Title);
2690 errs() << "ScheduleDAGMI::viewGraph is only available in debug builds on "
2691 << "systems with Graphviz or gv!\n";
2695 /// Out-of-line implementation with no arguments is handy for gdb.
2696 void ScheduleDAGMI::viewGraph() {
2697 viewGraph(getDAGName(), "Scheduling-Units Graph for " + getDAGName());