1 //===- MachineScheduler.cpp - Machine Instruction Scheduler ---------------===// 2 // 3 // The LLVM Compiler Infrastructure 4 // 5 // This file is distributed under the University of Illinois Open Source 6 // License. See LICENSE.TXT for details. 7 // 8 //===----------------------------------------------------------------------===// 9 // 10 // MachineScheduler schedules machine instructions after phi elimination. It 11 // preserves LiveIntervals so it can be invoked before register allocation. 12 // 13 //===----------------------------------------------------------------------===// 14 15 #define DEBUG_TYPE "misched" 16 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/Passes.h" 23 #include "llvm/CodeGen/RegisterClassInfo.h" 24 #include "llvm/CodeGen/ScheduleDFS.h" 25 #include "llvm/CodeGen/ScheduleHazardRecognizer.h" 26 #include "llvm/Support/CommandLine.h" 27 #include "llvm/Support/Debug.h" 28 #include "llvm/Support/ErrorHandling.h" 29 #include "llvm/Support/raw_ostream.h" 30 #include <queue> 31 32 using namespace llvm; 33 34 namespace llvm { 35 cl::opt<bool> ForceTopDown("misched-topdown", cl::Hidden, 36 cl::desc("Force top-down list scheduling")); 37 cl::opt<bool> ForceBottomUp("misched-bottomup", cl::Hidden, 38 cl::desc("Force bottom-up list scheduling")); 39 } 40 41 #ifndef NDEBUG 42 static cl::opt<bool> ViewMISchedDAGs("view-misched-dags", cl::Hidden, 43 cl::desc("Pop up a window to show MISched dags after they are processed")); 44 45 static cl::opt<unsigned> MISchedCutoff("misched-cutoff", cl::Hidden, 46 cl::desc("Stop scheduling after N instructions"), cl::init(~0U)); 47 #else 48 static bool ViewMISchedDAGs = false; 49 #endif // NDEBUG 50 51 // Experimental heuristics 52 static cl::opt<bool> EnableLoadCluster("misched-cluster", cl::Hidden, 53 cl::desc("Enable load clustering."), cl::init(true)); 54 55 // Experimental heuristics 56 static cl::opt<bool> EnableMacroFusion("misched-fusion", cl::Hidden, 57 cl::desc("Enable scheduling for macro fusion."), cl::init(true)); 58 59 //===----------------------------------------------------------------------===// 60 // Machine Instruction Scheduling Pass and Registry 61 //===----------------------------------------------------------------------===// 62 63 MachineSchedContext::MachineSchedContext(): 64 MF(0), MLI(0), MDT(0), PassConfig(0), AA(0), LIS(0) { 65 RegClassInfo = new RegisterClassInfo(); 66 } 67 68 MachineSchedContext::~MachineSchedContext() { 69 delete RegClassInfo; 70 } 71 72 namespace { 73 /// MachineScheduler runs after coalescing and before register allocation. 74 class MachineScheduler : public MachineSchedContext, 75 public MachineFunctionPass { 76 public: 77 MachineScheduler(); 78 79 virtual void getAnalysisUsage(AnalysisUsage &AU) const; 80 81 virtual void releaseMemory() {} 82 83 virtual bool runOnMachineFunction(MachineFunction&); 84 85 virtual void print(raw_ostream &O, const Module* = 0) const; 86 87 static char ID; // Class identification, replacement for typeinfo 88 }; 89 } // namespace 90 91 char MachineScheduler::ID = 0; 92 93 char &llvm::MachineSchedulerID = MachineScheduler::ID; 94 95 INITIALIZE_PASS_BEGIN(MachineScheduler, "misched", 96 "Machine Instruction Scheduler", false, false) 97 INITIALIZE_AG_DEPENDENCY(AliasAnalysis) 98 INITIALIZE_PASS_DEPENDENCY(SlotIndexes) 99 INITIALIZE_PASS_DEPENDENCY(LiveIntervals) 100 INITIALIZE_PASS_END(MachineScheduler, "misched", 101 "Machine Instruction Scheduler", false, false) 102 103 MachineScheduler::MachineScheduler() 104 : MachineFunctionPass(ID) { 105 initializeMachineSchedulerPass(*PassRegistry::getPassRegistry()); 106 } 107 108 void MachineScheduler::getAnalysisUsage(AnalysisUsage &AU) const { 109 AU.setPreservesCFG(); 110 AU.addRequiredID(MachineDominatorsID); 111 AU.addRequired<MachineLoopInfo>(); 112 AU.addRequired<AliasAnalysis>(); 113 AU.addRequired<TargetPassConfig>(); 114 AU.addRequired<SlotIndexes>(); 115 AU.addPreserved<SlotIndexes>(); 116 AU.addRequired<LiveIntervals>(); 117 AU.addPreserved<LiveIntervals>(); 118 MachineFunctionPass::getAnalysisUsage(AU); 119 } 120 121 MachinePassRegistry MachineSchedRegistry::Registry; 122 123 /// A dummy default scheduler factory indicates whether the scheduler 124 /// is overridden on the command line. 125 static ScheduleDAGInstrs *useDefaultMachineSched(MachineSchedContext *C) { 126 return 0; 127 } 128 129 /// MachineSchedOpt allows command line selection of the scheduler. 130 static cl::opt<MachineSchedRegistry::ScheduleDAGCtor, false, 131 RegisterPassParser<MachineSchedRegistry> > 132 MachineSchedOpt("misched", 133 cl::init(&useDefaultMachineSched), cl::Hidden, 134 cl::desc("Machine instruction scheduler to use")); 135 136 static MachineSchedRegistry 137 DefaultSchedRegistry("default", "Use the target's default scheduler choice.", 138 useDefaultMachineSched); 139 140 /// Forward declare the standard machine scheduler. This will be used as the 141 /// default scheduler if the target does not set a default. 142 static ScheduleDAGInstrs *createConvergingSched(MachineSchedContext *C); 143 144 145 /// Decrement this iterator until reaching the top or a non-debug instr. 146 static MachineBasicBlock::iterator 147 priorNonDebug(MachineBasicBlock::iterator I, MachineBasicBlock::iterator Beg) { 148 assert(I != Beg && "reached the top of the region, cannot decrement"); 149 while (--I != Beg) { 150 if (!I->isDebugValue()) 151 break; 152 } 153 return I; 154 } 155 156 /// If this iterator is a debug value, increment until reaching the End or a 157 /// non-debug instruction. 158 static MachineBasicBlock::iterator 159 nextIfDebug(MachineBasicBlock::iterator I, MachineBasicBlock::iterator End) { 160 for(; I != End; ++I) { 161 if (!I->isDebugValue()) 162 break; 163 } 164 return I; 165 } 166 167 /// Top-level MachineScheduler pass driver. 168 /// 169 /// Visit blocks in function order. Divide each block into scheduling regions 170 /// and visit them bottom-up. Visiting regions bottom-up is not required, but is 171 /// consistent with the DAG builder, which traverses the interior of the 172 /// scheduling regions bottom-up. 173 /// 174 /// This design avoids exposing scheduling boundaries to the DAG builder, 175 /// simplifying the DAG builder's support for "special" target instructions. 176 /// At the same time the design allows target schedulers to operate across 177 /// scheduling boundaries, for example to bundle the boudary instructions 178 /// without reordering them. This creates complexity, because the target 179 /// scheduler must update the RegionBegin and RegionEnd positions cached by 180 /// ScheduleDAGInstrs whenever adding or removing instructions. A much simpler 181 /// design would be to split blocks at scheduling boundaries, but LLVM has a 182 /// general bias against block splitting purely for implementation simplicity. 183 bool MachineScheduler::runOnMachineFunction(MachineFunction &mf) { 184 DEBUG(dbgs() << "Before MISsched:\n"; mf.print(dbgs())); 185 186 // Initialize the context of the pass. 187 MF = &mf; 188 MLI = &getAnalysis<MachineLoopInfo>(); 189 MDT = &getAnalysis<MachineDominatorTree>(); 190 PassConfig = &getAnalysis<TargetPassConfig>(); 191 AA = &getAnalysis<AliasAnalysis>(); 192 193 LIS = &getAnalysis<LiveIntervals>(); 194 const TargetInstrInfo *TII = MF->getTarget().getInstrInfo(); 195 196 RegClassInfo->runOnMachineFunction(*MF); 197 198 // Select the scheduler, or set the default. 199 MachineSchedRegistry::ScheduleDAGCtor Ctor = MachineSchedOpt; 200 if (Ctor == useDefaultMachineSched) { 201 // Get the default scheduler set by the target. 202 Ctor = MachineSchedRegistry::getDefault(); 203 if (!Ctor) { 204 Ctor = createConvergingSched; 205 MachineSchedRegistry::setDefault(Ctor); 206 } 207 } 208 // Instantiate the selected scheduler. 209 OwningPtr<ScheduleDAGInstrs> Scheduler(Ctor(this)); 210 211 // Visit all machine basic blocks. 212 // 213 // TODO: Visit blocks in global postorder or postorder within the bottom-up 214 // loop tree. Then we can optionally compute global RegPressure. 215 for (MachineFunction::iterator MBB = MF->begin(), MBBEnd = MF->end(); 216 MBB != MBBEnd; ++MBB) { 217 218 Scheduler->startBlock(MBB); 219 220 // Break the block into scheduling regions [I, RegionEnd), and schedule each 221 // region as soon as it is discovered. RegionEnd points the scheduling 222 // boundary at the bottom of the region. The DAG does not include RegionEnd, 223 // but the region does (i.e. the next RegionEnd is above the previous 224 // RegionBegin). If the current block has no terminator then RegionEnd == 225 // MBB->end() for the bottom region. 226 // 227 // The Scheduler may insert instructions during either schedule() or 228 // exitRegion(), even for empty regions. So the local iterators 'I' and 229 // 'RegionEnd' are invalid across these calls. 230 unsigned RemainingInstrs = MBB->size(); 231 for(MachineBasicBlock::iterator RegionEnd = MBB->end(); 232 RegionEnd != MBB->begin(); RegionEnd = Scheduler->begin()) { 233 234 // Avoid decrementing RegionEnd for blocks with no terminator. 235 if (RegionEnd != MBB->end() 236 || TII->isSchedulingBoundary(llvm::prior(RegionEnd), MBB, *MF)) { 237 --RegionEnd; 238 // Count the boundary instruction. 239 --RemainingInstrs; 240 } 241 242 // The next region starts above the previous region. Look backward in the 243 // instruction stream until we find the nearest boundary. 244 MachineBasicBlock::iterator I = RegionEnd; 245 for(;I != MBB->begin(); --I, --RemainingInstrs) { 246 if (TII->isSchedulingBoundary(llvm::prior(I), MBB, *MF)) 247 break; 248 } 249 // Notify the scheduler of the region, even if we may skip scheduling 250 // it. Perhaps it still needs to be bundled. 251 Scheduler->enterRegion(MBB, I, RegionEnd, RemainingInstrs); 252 253 // Skip empty scheduling regions (0 or 1 schedulable instructions). 254 if (I == RegionEnd || I == llvm::prior(RegionEnd)) { 255 // Close the current region. Bundle the terminator if needed. 256 // This invalidates 'RegionEnd' and 'I'. 257 Scheduler->exitRegion(); 258 continue; 259 } 260 DEBUG(dbgs() << "********** MI Scheduling **********\n"); 261 DEBUG(dbgs() << MF->getName() 262 << ":BB#" << MBB->getNumber() << "\n From: " << *I << " To: "; 263 if (RegionEnd != MBB->end()) dbgs() << *RegionEnd; 264 else dbgs() << "End"; 265 dbgs() << " Remaining: " << RemainingInstrs << "\n"); 266 267 // Schedule a region: possibly reorder instructions. 268 // This invalidates 'RegionEnd' and 'I'. 269 Scheduler->schedule(); 270 271 // Close the current region. 272 Scheduler->exitRegion(); 273 274 // Scheduling has invalidated the current iterator 'I'. Ask the 275 // scheduler for the top of it's scheduled region. 276 RegionEnd = Scheduler->begin(); 277 } 278 assert(RemainingInstrs == 0 && "Instruction count mismatch!"); 279 Scheduler->finishBlock(); 280 } 281 Scheduler->finalizeSchedule(); 282 DEBUG(LIS->print(dbgs())); 283 return true; 284 } 285 286 void MachineScheduler::print(raw_ostream &O, const Module* m) const { 287 // unimplemented 288 } 289 290 #if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP) 291 void ReadyQueue::dump() { 292 dbgs() << Name << ": "; 293 for (unsigned i = 0, e = Queue.size(); i < e; ++i) 294 dbgs() << Queue[i]->NodeNum << " "; 295 dbgs() << "\n"; 296 } 297 #endif 298 299 //===----------------------------------------------------------------------===// 300 // ScheduleDAGMI - Base class for MachineInstr scheduling with LiveIntervals 301 // preservation. 302 //===----------------------------------------------------------------------===// 303 304 bool ScheduleDAGMI::addEdge(SUnit *SuccSU, const SDep &PredDep) { 305 if (SuccSU != &ExitSU) { 306 // Do not use WillCreateCycle, it assumes SD scheduling. 307 // If Pred is reachable from Succ, then the edge creates a cycle. 308 if (Topo.IsReachable(PredDep.getSUnit(), SuccSU)) 309 return false; 310 Topo.AddPred(SuccSU, PredDep.getSUnit()); 311 } 312 SuccSU->addPred(PredDep, /*Required=*/!PredDep.isArtificial()); 313 // Return true regardless of whether a new edge needed to be inserted. 314 return true; 315 } 316 317 /// ReleaseSucc - Decrement the NumPredsLeft count of a successor. When 318 /// NumPredsLeft reaches zero, release the successor node. 319 /// 320 /// FIXME: Adjust SuccSU height based on MinLatency. 321 void ScheduleDAGMI::releaseSucc(SUnit *SU, SDep *SuccEdge) { 322 SUnit *SuccSU = SuccEdge->getSUnit(); 323 324 if (SuccEdge->isWeak()) { 325 --SuccSU->WeakPredsLeft; 326 if (SuccEdge->isCluster()) 327 NextClusterSucc = SuccSU; 328 return; 329 } 330 #ifndef NDEBUG 331 if (SuccSU->NumPredsLeft == 0) { 332 dbgs() << "*** Scheduling failed! ***\n"; 333 SuccSU->dump(this); 334 dbgs() << " has been released too many times!\n"; 335 llvm_unreachable(0); 336 } 337 #endif 338 --SuccSU->NumPredsLeft; 339 if (SuccSU->NumPredsLeft == 0 && SuccSU != &ExitSU) 340 SchedImpl->releaseTopNode(SuccSU); 341 } 342 343 /// releaseSuccessors - Call releaseSucc on each of SU's successors. 344 void ScheduleDAGMI::releaseSuccessors(SUnit *SU) { 345 for (SUnit::succ_iterator I = SU->Succs.begin(), E = SU->Succs.end(); 346 I != E; ++I) { 347 releaseSucc(SU, &*I); 348 } 349 } 350 351 /// ReleasePred - Decrement the NumSuccsLeft count of a predecessor. When 352 /// NumSuccsLeft reaches zero, release the predecessor node. 353 /// 354 /// FIXME: Adjust PredSU height based on MinLatency. 355 void ScheduleDAGMI::releasePred(SUnit *SU, SDep *PredEdge) { 356 SUnit *PredSU = PredEdge->getSUnit(); 357 358 if (PredEdge->isWeak()) { 359 --PredSU->WeakSuccsLeft; 360 if (PredEdge->isCluster()) 361 NextClusterPred = PredSU; 362 return; 363 } 364 #ifndef NDEBUG 365 if (PredSU->NumSuccsLeft == 0) { 366 dbgs() << "*** Scheduling failed! ***\n"; 367 PredSU->dump(this); 368 dbgs() << " has been released too many times!\n"; 369 llvm_unreachable(0); 370 } 371 #endif 372 --PredSU->NumSuccsLeft; 373 if (PredSU->NumSuccsLeft == 0 && PredSU != &EntrySU) 374 SchedImpl->releaseBottomNode(PredSU); 375 } 376 377 /// releasePredecessors - Call releasePred on each of SU's predecessors. 378 void ScheduleDAGMI::releasePredecessors(SUnit *SU) { 379 for (SUnit::pred_iterator I = SU->Preds.begin(), E = SU->Preds.end(); 380 I != E; ++I) { 381 releasePred(SU, &*I); 382 } 383 } 384 385 void ScheduleDAGMI::moveInstruction(MachineInstr *MI, 386 MachineBasicBlock::iterator InsertPos) { 387 // Advance RegionBegin if the first instruction moves down. 388 if (&*RegionBegin == MI) 389 ++RegionBegin; 390 391 // Update the instruction stream. 392 BB->splice(InsertPos, BB, MI); 393 394 // Update LiveIntervals 395 LIS->handleMove(MI, /*UpdateFlags=*/true); 396 397 // Recede RegionBegin if an instruction moves above the first. 398 if (RegionBegin == InsertPos) 399 RegionBegin = MI; 400 } 401 402 bool ScheduleDAGMI::checkSchedLimit() { 403 #ifndef NDEBUG 404 if (NumInstrsScheduled == MISchedCutoff && MISchedCutoff != ~0U) { 405 CurrentTop = CurrentBottom; 406 return false; 407 } 408 ++NumInstrsScheduled; 409 #endif 410 return true; 411 } 412 413 /// enterRegion - Called back from MachineScheduler::runOnMachineFunction after 414 /// crossing a scheduling boundary. [begin, end) includes all instructions in 415 /// the region, including the boundary itself and single-instruction regions 416 /// that don't get scheduled. 417 void ScheduleDAGMI::enterRegion(MachineBasicBlock *bb, 418 MachineBasicBlock::iterator begin, 419 MachineBasicBlock::iterator end, 420 unsigned endcount) 421 { 422 ScheduleDAGInstrs::enterRegion(bb, begin, end, endcount); 423 424 // For convenience remember the end of the liveness region. 425 LiveRegionEnd = 426 (RegionEnd == bb->end()) ? RegionEnd : llvm::next(RegionEnd); 427 } 428 429 // Setup the register pressure trackers for the top scheduled top and bottom 430 // scheduled regions. 431 void ScheduleDAGMI::initRegPressure() { 432 TopRPTracker.init(&MF, RegClassInfo, LIS, BB, RegionBegin); 433 BotRPTracker.init(&MF, RegClassInfo, LIS, BB, LiveRegionEnd); 434 435 // Close the RPTracker to finalize live ins. 436 RPTracker.closeRegion(); 437 438 DEBUG(RPTracker.getPressure().dump(TRI)); 439 440 // Initialize the live ins and live outs. 441 TopRPTracker.addLiveRegs(RPTracker.getPressure().LiveInRegs); 442 BotRPTracker.addLiveRegs(RPTracker.getPressure().LiveOutRegs); 443 444 // Close one end of the tracker so we can call 445 // getMaxUpward/DownwardPressureDelta before advancing across any 446 // instructions. This converts currently live regs into live ins/outs. 447 TopRPTracker.closeTop(); 448 BotRPTracker.closeBottom(); 449 450 // Account for liveness generated by the region boundary. 451 if (LiveRegionEnd != RegionEnd) 452 BotRPTracker.recede(); 453 454 assert(BotRPTracker.getPos() == RegionEnd && "Can't find the region bottom"); 455 456 // Cache the list of excess pressure sets in this region. This will also track 457 // the max pressure in the scheduled code for these sets. 458 RegionCriticalPSets.clear(); 459 std::vector<unsigned> RegionPressure = RPTracker.getPressure().MaxSetPressure; 460 for (unsigned i = 0, e = RegionPressure.size(); i < e; ++i) { 461 unsigned Limit = TRI->getRegPressureSetLimit(i); 462 DEBUG(dbgs() << TRI->getRegPressureSetName(i) 463 << "Limit " << Limit 464 << " Actual " << RegionPressure[i] << "\n"); 465 if (RegionPressure[i] > Limit) 466 RegionCriticalPSets.push_back(PressureElement(i, 0)); 467 } 468 DEBUG(dbgs() << "Excess PSets: "; 469 for (unsigned i = 0, e = RegionCriticalPSets.size(); i != e; ++i) 470 dbgs() << TRI->getRegPressureSetName( 471 RegionCriticalPSets[i].PSetID) << " "; 472 dbgs() << "\n"); 473 } 474 475 // FIXME: When the pressure tracker deals in pressure differences then we won't 476 // iterate over all RegionCriticalPSets[i]. 477 void ScheduleDAGMI:: 478 updateScheduledPressure(std::vector<unsigned> NewMaxPressure) { 479 for (unsigned i = 0, e = RegionCriticalPSets.size(); i < e; ++i) { 480 unsigned ID = RegionCriticalPSets[i].PSetID; 481 int &MaxUnits = RegionCriticalPSets[i].UnitIncrease; 482 if ((int)NewMaxPressure[ID] > MaxUnits) 483 MaxUnits = NewMaxPressure[ID]; 484 } 485 } 486 487 /// schedule - Called back from MachineScheduler::runOnMachineFunction 488 /// after setting up the current scheduling region. [RegionBegin, RegionEnd) 489 /// only includes instructions that have DAG nodes, not scheduling boundaries. 490 /// 491 /// This is a skeletal driver, with all the functionality pushed into helpers, 492 /// so that it can be easilly extended by experimental schedulers. Generally, 493 /// implementing MachineSchedStrategy should be sufficient to implement a new 494 /// scheduling algorithm. However, if a scheduler further subclasses 495 /// ScheduleDAGMI then it will want to override this virtual method in order to 496 /// update any specialized state. 497 void ScheduleDAGMI::schedule() { 498 buildDAGWithRegPressure(); 499 500 Topo.InitDAGTopologicalSorting(); 501 502 postprocessDAG(); 503 504 DEBUG(for (unsigned su = 0, e = SUnits.size(); su != e; ++su) 505 SUnits[su].dumpAll(this)); 506 507 if (ViewMISchedDAGs) viewGraph(); 508 509 initQueues(); 510 511 bool IsTopNode = false; 512 while (SUnit *SU = SchedImpl->pickNode(IsTopNode)) { 513 assert(!SU->isScheduled && "Node already scheduled"); 514 if (!checkSchedLimit()) 515 break; 516 517 scheduleMI(SU, IsTopNode); 518 519 updateQueues(SU, IsTopNode); 520 } 521 assert(CurrentTop == CurrentBottom && "Nonempty unscheduled zone."); 522 523 placeDebugValues(); 524 525 DEBUG({ 526 unsigned BBNum = begin()->getParent()->getNumber(); 527 dbgs() << "*** Final schedule for BB#" << BBNum << " ***\n"; 528 dumpSchedule(); 529 dbgs() << '\n'; 530 }); 531 } 532 533 /// Build the DAG and setup three register pressure trackers. 534 void ScheduleDAGMI::buildDAGWithRegPressure() { 535 // Initialize the register pressure tracker used by buildSchedGraph. 536 RPTracker.init(&MF, RegClassInfo, LIS, BB, LiveRegionEnd); 537 538 // Account for liveness generate by the region boundary. 539 if (LiveRegionEnd != RegionEnd) 540 RPTracker.recede(); 541 542 // Build the DAG, and compute current register pressure. 543 buildSchedGraph(AA, &RPTracker); 544 if (ViewMISchedDAGs) viewGraph(); 545 546 // Initialize top/bottom trackers after computing region pressure. 547 initRegPressure(); 548 } 549 550 /// Apply each ScheduleDAGMutation step in order. 551 void ScheduleDAGMI::postprocessDAG() { 552 for (unsigned i = 0, e = Mutations.size(); i < e; ++i) { 553 Mutations[i]->apply(this); 554 } 555 } 556 557 // Release all DAG roots for scheduling. 558 // 559 // Nodes with unreleased weak edges can still be roots. 560 void ScheduleDAGMI::releaseRoots() { 561 SmallVector<SUnit*, 16> BotRoots; 562 563 for (std::vector<SUnit>::iterator 564 I = SUnits.begin(), E = SUnits.end(); I != E; ++I) { 565 SUnit *SU = &(*I); 566 // A SUnit is ready to top schedule if it has no predecessors. 567 if (!I->NumPredsLeft && SU != &EntrySU) 568 SchedImpl->releaseTopNode(SU); 569 // A SUnit is ready to bottom schedule if it has no successors. 570 if (!I->NumSuccsLeft && SU != &ExitSU) 571 BotRoots.push_back(SU); 572 } 573 // Release bottom roots in reverse order so the higher priority nodes appear 574 // first. This is more natural and slightly more efficient. 575 for (SmallVectorImpl<SUnit*>::const_reverse_iterator 576 I = BotRoots.rbegin(), E = BotRoots.rend(); I != E; ++I) 577 SchedImpl->releaseBottomNode(*I); 578 } 579 580 /// Identify DAG roots and setup scheduler queues. 581 void ScheduleDAGMI::initQueues() { 582 NextClusterSucc = NULL; 583 NextClusterPred = NULL; 584 585 // Initialize the strategy before modifying the DAG. 586 SchedImpl->initialize(this); 587 588 // Release all DAG roots for scheduling, not including EntrySU/ExitSU. 589 releaseRoots(); 590 591 releaseSuccessors(&EntrySU); 592 releasePredecessors(&ExitSU); 593 594 SchedImpl->registerRoots(); 595 596 // Advance past initial DebugValues. 597 assert(TopRPTracker.getPos() == RegionBegin && "bad initial Top tracker"); 598 CurrentTop = nextIfDebug(RegionBegin, RegionEnd); 599 TopRPTracker.setPos(CurrentTop); 600 601 CurrentBottom = RegionEnd; 602 } 603 604 /// Move an instruction and update register pressure. 605 void ScheduleDAGMI::scheduleMI(SUnit *SU, bool IsTopNode) { 606 // Move the instruction to its new location in the instruction stream. 607 MachineInstr *MI = SU->getInstr(); 608 609 if (IsTopNode) { 610 assert(SU->isTopReady() && "node still has unscheduled dependencies"); 611 if (&*CurrentTop == MI) 612 CurrentTop = nextIfDebug(++CurrentTop, CurrentBottom); 613 else { 614 moveInstruction(MI, CurrentTop); 615 TopRPTracker.setPos(MI); 616 } 617 618 // Update top scheduled pressure. 619 TopRPTracker.advance(); 620 assert(TopRPTracker.getPos() == CurrentTop && "out of sync"); 621 updateScheduledPressure(TopRPTracker.getPressure().MaxSetPressure); 622 } 623 else { 624 assert(SU->isBottomReady() && "node still has unscheduled dependencies"); 625 MachineBasicBlock::iterator priorII = 626 priorNonDebug(CurrentBottom, CurrentTop); 627 if (&*priorII == MI) 628 CurrentBottom = priorII; 629 else { 630 if (&*CurrentTop == MI) { 631 CurrentTop = nextIfDebug(++CurrentTop, priorII); 632 TopRPTracker.setPos(CurrentTop); 633 } 634 moveInstruction(MI, CurrentBottom); 635 CurrentBottom = MI; 636 } 637 // Update bottom scheduled pressure. 638 BotRPTracker.recede(); 639 assert(BotRPTracker.getPos() == CurrentBottom && "out of sync"); 640 updateScheduledPressure(BotRPTracker.getPressure().MaxSetPressure); 641 } 642 } 643 644 /// Update scheduler queues after scheduling an instruction. 645 void ScheduleDAGMI::updateQueues(SUnit *SU, bool IsTopNode) { 646 // Release dependent instructions for scheduling. 647 if (IsTopNode) 648 releaseSuccessors(SU); 649 else 650 releasePredecessors(SU); 651 652 SU->isScheduled = true; 653 654 // Notify the scheduling strategy after updating the DAG. 655 SchedImpl->schedNode(SU, IsTopNode); 656 } 657 658 /// Reinsert any remaining debug_values, just like the PostRA scheduler. 659 void ScheduleDAGMI::placeDebugValues() { 660 // If first instruction was a DBG_VALUE then put it back. 661 if (FirstDbgValue) { 662 BB->splice(RegionBegin, BB, FirstDbgValue); 663 RegionBegin = FirstDbgValue; 664 } 665 666 for (std::vector<std::pair<MachineInstr *, MachineInstr *> >::iterator 667 DI = DbgValues.end(), DE = DbgValues.begin(); DI != DE; --DI) { 668 std::pair<MachineInstr *, MachineInstr *> P = *prior(DI); 669 MachineInstr *DbgValue = P.first; 670 MachineBasicBlock::iterator OrigPrevMI = P.second; 671 if (&*RegionBegin == DbgValue) 672 ++RegionBegin; 673 BB->splice(++OrigPrevMI, BB, DbgValue); 674 if (OrigPrevMI == llvm::prior(RegionEnd)) 675 RegionEnd = DbgValue; 676 } 677 DbgValues.clear(); 678 FirstDbgValue = NULL; 679 } 680 681 #if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP) 682 void ScheduleDAGMI::dumpSchedule() const { 683 for (MachineBasicBlock::iterator MI = begin(), ME = end(); MI != ME; ++MI) { 684 if (SUnit *SU = getSUnit(&(*MI))) 685 SU->dump(this); 686 else 687 dbgs() << "Missing SUnit\n"; 688 } 689 } 690 #endif 691 692 //===----------------------------------------------------------------------===// 693 // LoadClusterMutation - DAG post-processing to cluster loads. 694 //===----------------------------------------------------------------------===// 695 696 namespace { 697 /// \brief Post-process the DAG to create cluster edges between neighboring 698 /// loads. 699 class LoadClusterMutation : public ScheduleDAGMutation { 700 struct LoadInfo { 701 SUnit *SU; 702 unsigned BaseReg; 703 unsigned Offset; 704 LoadInfo(SUnit *su, unsigned reg, unsigned ofs) 705 : SU(su), BaseReg(reg), Offset(ofs) {} 706 }; 707 static bool LoadInfoLess(const LoadClusterMutation::LoadInfo &LHS, 708 const LoadClusterMutation::LoadInfo &RHS); 709 710 const TargetInstrInfo *TII; 711 const TargetRegisterInfo *TRI; 712 public: 713 LoadClusterMutation(const TargetInstrInfo *tii, 714 const TargetRegisterInfo *tri) 715 : TII(tii), TRI(tri) {} 716 717 virtual void apply(ScheduleDAGMI *DAG); 718 protected: 719 void clusterNeighboringLoads(ArrayRef<SUnit*> Loads, ScheduleDAGMI *DAG); 720 }; 721 } // anonymous 722 723 bool LoadClusterMutation::LoadInfoLess( 724 const LoadClusterMutation::LoadInfo &LHS, 725 const LoadClusterMutation::LoadInfo &RHS) { 726 if (LHS.BaseReg != RHS.BaseReg) 727 return LHS.BaseReg < RHS.BaseReg; 728 return LHS.Offset < RHS.Offset; 729 } 730 731 void LoadClusterMutation::clusterNeighboringLoads(ArrayRef<SUnit*> Loads, 732 ScheduleDAGMI *DAG) { 733 SmallVector<LoadClusterMutation::LoadInfo,32> LoadRecords; 734 for (unsigned Idx = 0, End = Loads.size(); Idx != End; ++Idx) { 735 SUnit *SU = Loads[Idx]; 736 unsigned BaseReg; 737 unsigned Offset; 738 if (TII->getLdStBaseRegImmOfs(SU->getInstr(), BaseReg, Offset, TRI)) 739 LoadRecords.push_back(LoadInfo(SU, BaseReg, Offset)); 740 } 741 if (LoadRecords.size() < 2) 742 return; 743 std::sort(LoadRecords.begin(), LoadRecords.end(), LoadInfoLess); 744 unsigned ClusterLength = 1; 745 for (unsigned Idx = 0, End = LoadRecords.size(); Idx < (End - 1); ++Idx) { 746 if (LoadRecords[Idx].BaseReg != LoadRecords[Idx+1].BaseReg) { 747 ClusterLength = 1; 748 continue; 749 } 750 751 SUnit *SUa = LoadRecords[Idx].SU; 752 SUnit *SUb = LoadRecords[Idx+1].SU; 753 if (TII->shouldClusterLoads(SUa->getInstr(), SUb->getInstr(), ClusterLength) 754 && DAG->addEdge(SUb, SDep(SUa, SDep::Cluster))) { 755 756 DEBUG(dbgs() << "Cluster loads SU(" << SUa->NodeNum << ") - SU(" 757 << SUb->NodeNum << ")\n"); 758 // Copy successor edges from SUa to SUb. Interleaving computation 759 // dependent on SUa can prevent load combining due to register reuse. 760 // Predecessor edges do not need to be copied from SUb to SUa since nearby 761 // loads should have effectively the same inputs. 762 for (SUnit::const_succ_iterator 763 SI = SUa->Succs.begin(), SE = SUa->Succs.end(); SI != SE; ++SI) { 764 if (SI->getSUnit() == SUb) 765 continue; 766 DEBUG(dbgs() << " Copy Succ SU(" << SI->getSUnit()->NodeNum << ")\n"); 767 DAG->addEdge(SI->getSUnit(), SDep(SUb, SDep::Artificial)); 768 } 769 ++ClusterLength; 770 } 771 else 772 ClusterLength = 1; 773 } 774 } 775 776 /// \brief Callback from DAG postProcessing to create cluster edges for loads. 777 void LoadClusterMutation::apply(ScheduleDAGMI *DAG) { 778 // Map DAG NodeNum to store chain ID. 779 DenseMap<unsigned, unsigned> StoreChainIDs; 780 // Map each store chain to a set of dependent loads. 781 SmallVector<SmallVector<SUnit*,4>, 32> StoreChainDependents; 782 for (unsigned Idx = 0, End = DAG->SUnits.size(); Idx != End; ++Idx) { 783 SUnit *SU = &DAG->SUnits[Idx]; 784 if (!SU->getInstr()->mayLoad()) 785 continue; 786 unsigned ChainPredID = DAG->SUnits.size(); 787 for (SUnit::const_pred_iterator 788 PI = SU->Preds.begin(), PE = SU->Preds.end(); PI != PE; ++PI) { 789 if (PI->isCtrl()) { 790 ChainPredID = PI->getSUnit()->NodeNum; 791 break; 792 } 793 } 794 // Check if this chain-like pred has been seen 795 // before. ChainPredID==MaxNodeID for loads at the top of the schedule. 796 unsigned NumChains = StoreChainDependents.size(); 797 std::pair<DenseMap<unsigned, unsigned>::iterator, bool> Result = 798 StoreChainIDs.insert(std::make_pair(ChainPredID, NumChains)); 799 if (Result.second) 800 StoreChainDependents.resize(NumChains + 1); 801 StoreChainDependents[Result.first->second].push_back(SU); 802 } 803 // Iterate over the store chains. 804 for (unsigned Idx = 0, End = StoreChainDependents.size(); Idx != End; ++Idx) 805 clusterNeighboringLoads(StoreChainDependents[Idx], DAG); 806 } 807 808 //===----------------------------------------------------------------------===// 809 // MacroFusion - DAG post-processing to encourage fusion of macro ops. 810 //===----------------------------------------------------------------------===// 811 812 namespace { 813 /// \brief Post-process the DAG to create cluster edges between instructions 814 /// that may be fused by the processor into a single operation. 815 class MacroFusion : public ScheduleDAGMutation { 816 const TargetInstrInfo *TII; 817 public: 818 MacroFusion(const TargetInstrInfo *tii): TII(tii) {} 819 820 virtual void apply(ScheduleDAGMI *DAG); 821 }; 822 } // anonymous 823 824 /// \brief Callback from DAG postProcessing to create cluster edges to encourage 825 /// fused operations. 826 void MacroFusion::apply(ScheduleDAGMI *DAG) { 827 // For now, assume targets can only fuse with the branch. 828 MachineInstr *Branch = DAG->ExitSU.getInstr(); 829 if (!Branch) 830 return; 831 832 for (unsigned Idx = DAG->SUnits.size(); Idx > 0;) { 833 SUnit *SU = &DAG->SUnits[--Idx]; 834 if (!TII->shouldScheduleAdjacent(SU->getInstr(), Branch)) 835 continue; 836 837 // Create a single weak edge from SU to ExitSU. The only effect is to cause 838 // bottom-up scheduling to heavily prioritize the clustered SU. There is no 839 // need to copy predecessor edges from ExitSU to SU, since top-down 840 // scheduling cannot prioritize ExitSU anyway. To defer top-down scheduling 841 // of SU, we could create an artificial edge from the deepest root, but it 842 // hasn't been needed yet. 843 bool Success = DAG->addEdge(&DAG->ExitSU, SDep(SU, SDep::Cluster)); 844 (void)Success; 845 assert(Success && "No DAG nodes should be reachable from ExitSU"); 846 847 DEBUG(dbgs() << "Macro Fuse SU(" << SU->NodeNum << ")\n"); 848 break; 849 } 850 } 851 852 //===----------------------------------------------------------------------===// 853 // ConvergingScheduler - Implementation of the standard MachineSchedStrategy. 854 //===----------------------------------------------------------------------===// 855 856 namespace { 857 /// ConvergingScheduler shrinks the unscheduled zone using heuristics to balance 858 /// the schedule. 859 class ConvergingScheduler : public MachineSchedStrategy { 860 public: 861 /// Represent the type of SchedCandidate found within a single queue. 862 /// pickNodeBidirectional depends on these listed by decreasing priority. 863 enum CandReason { 864 NoCand, SingleExcess, SingleCritical, Cluster, 865 ResourceReduce, ResourceDemand, BotHeightReduce, BotPathReduce, 866 TopDepthReduce, TopPathReduce, SingleMax, MultiPressure, NextDefUse, 867 NodeOrder}; 868 869 #ifndef NDEBUG 870 static const char *getReasonStr(ConvergingScheduler::CandReason Reason); 871 #endif 872 873 /// Policy for scheduling the next instruction in the candidate's zone. 874 struct CandPolicy { 875 bool ReduceLatency; 876 unsigned ReduceResIdx; 877 unsigned DemandResIdx; 878 879 CandPolicy(): ReduceLatency(false), ReduceResIdx(0), DemandResIdx(0) {} 880 }; 881 882 /// Status of an instruction's critical resource consumption. 883 struct SchedResourceDelta { 884 // Count critical resources in the scheduled region required by SU. 885 unsigned CritResources; 886 887 // Count critical resources from another region consumed by SU. 888 unsigned DemandedResources; 889 890 SchedResourceDelta(): CritResources(0), DemandedResources(0) {} 891 892 bool operator==(const SchedResourceDelta &RHS) const { 893 return CritResources == RHS.CritResources 894 && DemandedResources == RHS.DemandedResources; 895 } 896 bool operator!=(const SchedResourceDelta &RHS) const { 897 return !operator==(RHS); 898 } 899 }; 900 901 /// Store the state used by ConvergingScheduler heuristics, required for the 902 /// lifetime of one invocation of pickNode(). 903 struct SchedCandidate { 904 CandPolicy Policy; 905 906 // The best SUnit candidate. 907 SUnit *SU; 908 909 // The reason for this candidate. 910 CandReason Reason; 911 912 // Register pressure values for the best candidate. 913 RegPressureDelta RPDelta; 914 915 // Critical resource consumption of the best candidate. 916 SchedResourceDelta ResDelta; 917 918 SchedCandidate(const CandPolicy &policy) 919 : Policy(policy), SU(NULL), Reason(NoCand) {} 920 921 bool isValid() const { return SU; } 922 923 // Copy the status of another candidate without changing policy. 924 void setBest(SchedCandidate &Best) { 925 assert(Best.Reason != NoCand && "uninitialized Sched candidate"); 926 SU = Best.SU; 927 Reason = Best.Reason; 928 RPDelta = Best.RPDelta; 929 ResDelta = Best.ResDelta; 930 } 931 932 void initResourceDelta(const ScheduleDAGMI *DAG, 933 const TargetSchedModel *SchedModel); 934 }; 935 936 /// Summarize the unscheduled region. 937 struct SchedRemainder { 938 // Critical path through the DAG in expected latency. 939 unsigned CriticalPath; 940 941 // Unscheduled resources 942 SmallVector<unsigned, 16> RemainingCounts; 943 // Critical resource for the unscheduled zone. 944 unsigned CritResIdx; 945 // Number of micro-ops left to schedule. 946 unsigned RemainingMicroOps; 947 948 void reset() { 949 CriticalPath = 0; 950 RemainingCounts.clear(); 951 CritResIdx = 0; 952 RemainingMicroOps = 0; 953 } 954 955 SchedRemainder() { reset(); } 956 957 void init(ScheduleDAGMI *DAG, const TargetSchedModel *SchedModel); 958 959 unsigned getMaxRemainingCount(const TargetSchedModel *SchedModel) const { 960 if (!SchedModel->hasInstrSchedModel()) 961 return 0; 962 963 return std::max( 964 RemainingMicroOps * SchedModel->getMicroOpFactor(), 965 RemainingCounts[CritResIdx]); 966 } 967 }; 968 969 /// Each Scheduling boundary is associated with ready queues. It tracks the 970 /// current cycle in the direction of movement, and maintains the state 971 /// of "hazards" and other interlocks at the current cycle. 972 struct SchedBoundary { 973 ScheduleDAGMI *DAG; 974 const TargetSchedModel *SchedModel; 975 SchedRemainder *Rem; 976 977 ReadyQueue Available; 978 ReadyQueue Pending; 979 bool CheckPending; 980 981 // For heuristics, keep a list of the nodes that immediately depend on the 982 // most recently scheduled node. 983 SmallPtrSet<const SUnit*, 8> NextSUs; 984 985 ScheduleHazardRecognizer *HazardRec; 986 987 unsigned CurrCycle; 988 unsigned IssueCount; 989 990 /// MinReadyCycle - Cycle of the soonest available instruction. 991 unsigned MinReadyCycle; 992 993 // The expected latency of the critical path in this scheduled zone. 994 unsigned ExpectedLatency; 995 996 // Resources used in the scheduled zone beyond this boundary. 997 SmallVector<unsigned, 16> ResourceCounts; 998 999 // Cache the critical resources ID in this scheduled zone. 1000 unsigned CritResIdx; 1001 1002 // Is the scheduled region resource limited vs. latency limited. 1003 bool IsResourceLimited; 1004 1005 unsigned ExpectedCount; 1006 1007 #ifndef NDEBUG 1008 // Remember the greatest min operand latency. 1009 unsigned MaxMinLatency; 1010 #endif 1011 1012 void reset() { 1013 Available.clear(); 1014 Pending.clear(); 1015 CheckPending = false; 1016 NextSUs.clear(); 1017 HazardRec = 0; 1018 CurrCycle = 0; 1019 IssueCount = 0; 1020 MinReadyCycle = UINT_MAX; 1021 ExpectedLatency = 0; 1022 ResourceCounts.resize(1); 1023 assert(!ResourceCounts[0] && "nonzero count for bad resource"); 1024 CritResIdx = 0; 1025 IsResourceLimited = false; 1026 ExpectedCount = 0; 1027 #ifndef NDEBUG 1028 MaxMinLatency = 0; 1029 #endif 1030 // Reserve a zero-count for invalid CritResIdx. 1031 ResourceCounts.resize(1); 1032 } 1033 1034 /// Pending queues extend the ready queues with the same ID and the 1035 /// PendingFlag set. 1036 SchedBoundary(unsigned ID, const Twine &Name): 1037 DAG(0), SchedModel(0), Rem(0), Available(ID, Name+".A"), 1038 Pending(ID << ConvergingScheduler::LogMaxQID, Name+".P") { 1039 reset(); 1040 } 1041 1042 ~SchedBoundary() { delete HazardRec; } 1043 1044 void init(ScheduleDAGMI *dag, const TargetSchedModel *smodel, 1045 SchedRemainder *rem); 1046 1047 bool isTop() const { 1048 return Available.getID() == ConvergingScheduler::TopQID; 1049 } 1050 1051 unsigned getUnscheduledLatency(SUnit *SU) const { 1052 if (isTop()) 1053 return SU->getHeight(); 1054 return SU->getDepth() + SU->Latency; 1055 } 1056 1057 unsigned getCriticalCount() const { 1058 return ResourceCounts[CritResIdx]; 1059 } 1060 1061 bool checkHazard(SUnit *SU); 1062 1063 void setLatencyPolicy(CandPolicy &Policy); 1064 1065 void releaseNode(SUnit *SU, unsigned ReadyCycle); 1066 1067 void bumpCycle(); 1068 1069 void countResource(unsigned PIdx, unsigned Cycles); 1070 1071 void bumpNode(SUnit *SU); 1072 1073 void releasePending(); 1074 1075 void removeReady(SUnit *SU); 1076 1077 SUnit *pickOnlyChoice(); 1078 }; 1079 1080 private: 1081 ScheduleDAGMI *DAG; 1082 const TargetSchedModel *SchedModel; 1083 const TargetRegisterInfo *TRI; 1084 1085 // State of the top and bottom scheduled instruction boundaries. 1086 SchedRemainder Rem; 1087 SchedBoundary Top; 1088 SchedBoundary Bot; 1089 1090 public: 1091 /// SUnit::NodeQueueId: 0 (none), 1 (top), 2 (bot), 3 (both) 1092 enum { 1093 TopQID = 1, 1094 BotQID = 2, 1095 LogMaxQID = 2 1096 }; 1097 1098 ConvergingScheduler(): 1099 DAG(0), SchedModel(0), TRI(0), Top(TopQID, "TopQ"), Bot(BotQID, "BotQ") {} 1100 1101 virtual void initialize(ScheduleDAGMI *dag); 1102 1103 virtual SUnit *pickNode(bool &IsTopNode); 1104 1105 virtual void schedNode(SUnit *SU, bool IsTopNode); 1106 1107 virtual void releaseTopNode(SUnit *SU); 1108 1109 virtual void releaseBottomNode(SUnit *SU); 1110 1111 virtual void registerRoots(); 1112 1113 protected: 1114 void balanceZones( 1115 ConvergingScheduler::SchedBoundary &CriticalZone, 1116 ConvergingScheduler::SchedCandidate &CriticalCand, 1117 ConvergingScheduler::SchedBoundary &OppositeZone, 1118 ConvergingScheduler::SchedCandidate &OppositeCand); 1119 1120 void checkResourceLimits(ConvergingScheduler::SchedCandidate &TopCand, 1121 ConvergingScheduler::SchedCandidate &BotCand); 1122 1123 void tryCandidate(SchedCandidate &Cand, 1124 SchedCandidate &TryCand, 1125 SchedBoundary &Zone, 1126 const RegPressureTracker &RPTracker, 1127 RegPressureTracker &TempTracker); 1128 1129 SUnit *pickNodeBidirectional(bool &IsTopNode); 1130 1131 void pickNodeFromQueue(SchedBoundary &Zone, 1132 const RegPressureTracker &RPTracker, 1133 SchedCandidate &Candidate); 1134 1135 #ifndef NDEBUG 1136 void traceCandidate(const SchedCandidate &Cand, const SchedBoundary &Zone); 1137 #endif 1138 }; 1139 } // namespace 1140 1141 void ConvergingScheduler::SchedRemainder:: 1142 init(ScheduleDAGMI *DAG, const TargetSchedModel *SchedModel) { 1143 reset(); 1144 if (!SchedModel->hasInstrSchedModel()) 1145 return; 1146 RemainingCounts.resize(SchedModel->getNumProcResourceKinds()); 1147 for (std::vector<SUnit>::iterator 1148 I = DAG->SUnits.begin(), E = DAG->SUnits.end(); I != E; ++I) { 1149 const MCSchedClassDesc *SC = DAG->getSchedClass(&*I); 1150 RemainingMicroOps += SchedModel->getNumMicroOps(I->getInstr(), SC); 1151 for (TargetSchedModel::ProcResIter 1152 PI = SchedModel->getWriteProcResBegin(SC), 1153 PE = SchedModel->getWriteProcResEnd(SC); PI != PE; ++PI) { 1154 unsigned PIdx = PI->ProcResourceIdx; 1155 unsigned Factor = SchedModel->getResourceFactor(PIdx); 1156 RemainingCounts[PIdx] += (Factor * PI->Cycles); 1157 } 1158 } 1159 for (unsigned PIdx = 0, PEnd = SchedModel->getNumProcResourceKinds(); 1160 PIdx != PEnd; ++PIdx) { 1161 if ((int)(RemainingCounts[PIdx] - RemainingCounts[CritResIdx]) 1162 >= (int)SchedModel->getLatencyFactor()) { 1163 CritResIdx = PIdx; 1164 } 1165 } 1166 } 1167 1168 void ConvergingScheduler::SchedBoundary:: 1169 init(ScheduleDAGMI *dag, const TargetSchedModel *smodel, SchedRemainder *rem) { 1170 reset(); 1171 DAG = dag; 1172 SchedModel = smodel; 1173 Rem = rem; 1174 if (SchedModel->hasInstrSchedModel()) 1175 ResourceCounts.resize(SchedModel->getNumProcResourceKinds()); 1176 } 1177 1178 void ConvergingScheduler::initialize(ScheduleDAGMI *dag) { 1179 DAG = dag; 1180 SchedModel = DAG->getSchedModel(); 1181 TRI = DAG->TRI; 1182 Rem.init(DAG, SchedModel); 1183 Top.init(DAG, SchedModel, &Rem); 1184 Bot.init(DAG, SchedModel, &Rem); 1185 1186 // Initialize resource counts. 1187 1188 // Initialize the HazardRecognizers. If itineraries don't exist, are empty, or 1189 // are disabled, then these HazardRecs will be disabled. 1190 const InstrItineraryData *Itin = SchedModel->getInstrItineraries(); 1191 const TargetMachine &TM = DAG->MF.getTarget(); 1192 Top.HazardRec = TM.getInstrInfo()->CreateTargetMIHazardRecognizer(Itin, DAG); 1193 Bot.HazardRec = TM.getInstrInfo()->CreateTargetMIHazardRecognizer(Itin, DAG); 1194 1195 assert((!ForceTopDown || !ForceBottomUp) && 1196 "-misched-topdown incompatible with -misched-bottomup"); 1197 } 1198 1199 void ConvergingScheduler::releaseTopNode(SUnit *SU) { 1200 if (SU->isScheduled) 1201 return; 1202 1203 for (SUnit::pred_iterator I = SU->Preds.begin(), E = SU->Preds.end(); 1204 I != E; ++I) { 1205 unsigned PredReadyCycle = I->getSUnit()->TopReadyCycle; 1206 unsigned MinLatency = I->getMinLatency(); 1207 #ifndef NDEBUG 1208 Top.MaxMinLatency = std::max(MinLatency, Top.MaxMinLatency); 1209 #endif 1210 if (SU->TopReadyCycle < PredReadyCycle + MinLatency) 1211 SU->TopReadyCycle = PredReadyCycle + MinLatency; 1212 } 1213 Top.releaseNode(SU, SU->TopReadyCycle); 1214 } 1215 1216 void ConvergingScheduler::releaseBottomNode(SUnit *SU) { 1217 if (SU->isScheduled) 1218 return; 1219 1220 assert(SU->getInstr() && "Scheduled SUnit must have instr"); 1221 1222 for (SUnit::succ_iterator I = SU->Succs.begin(), E = SU->Succs.end(); 1223 I != E; ++I) { 1224 if (I->isWeak()) 1225 continue; 1226 unsigned SuccReadyCycle = I->getSUnit()->BotReadyCycle; 1227 unsigned MinLatency = I->getMinLatency(); 1228 #ifndef NDEBUG 1229 Bot.MaxMinLatency = std::max(MinLatency, Bot.MaxMinLatency); 1230 #endif 1231 if (SU->BotReadyCycle < SuccReadyCycle + MinLatency) 1232 SU->BotReadyCycle = SuccReadyCycle + MinLatency; 1233 } 1234 Bot.releaseNode(SU, SU->BotReadyCycle); 1235 } 1236 1237 void ConvergingScheduler::registerRoots() { 1238 Rem.CriticalPath = DAG->ExitSU.getDepth(); 1239 // Some roots may not feed into ExitSU. Check all of them in case. 1240 for (std::vector<SUnit*>::const_iterator 1241 I = Bot.Available.begin(), E = Bot.Available.end(); I != E; ++I) { 1242 if ((*I)->getDepth() > Rem.CriticalPath) 1243 Rem.CriticalPath = (*I)->getDepth(); 1244 } 1245 DEBUG(dbgs() << "Critical Path: " << Rem.CriticalPath << '\n'); 1246 } 1247 1248 /// Does this SU have a hazard within the current instruction group. 1249 /// 1250 /// The scheduler supports two modes of hazard recognition. The first is the 1251 /// ScheduleHazardRecognizer API. It is a fully general hazard recognizer that 1252 /// supports highly complicated in-order reservation tables 1253 /// (ScoreboardHazardRecognizer) and arbitraty target-specific logic. 1254 /// 1255 /// The second is a streamlined mechanism that checks for hazards based on 1256 /// simple counters that the scheduler itself maintains. It explicitly checks 1257 /// for instruction dispatch limitations, including the number of micro-ops that 1258 /// can dispatch per cycle. 1259 /// 1260 /// TODO: Also check whether the SU must start a new group. 1261 bool ConvergingScheduler::SchedBoundary::checkHazard(SUnit *SU) { 1262 if (HazardRec->isEnabled()) 1263 return HazardRec->getHazardType(SU) != ScheduleHazardRecognizer::NoHazard; 1264 1265 unsigned uops = SchedModel->getNumMicroOps(SU->getInstr()); 1266 if ((IssueCount > 0) && (IssueCount + uops > SchedModel->getIssueWidth())) { 1267 DEBUG(dbgs() << " SU(" << SU->NodeNum << ") uops=" 1268 << SchedModel->getNumMicroOps(SU->getInstr()) << '\n'); 1269 return true; 1270 } 1271 return false; 1272 } 1273 1274 /// Compute the remaining latency to determine whether ILP should be increased. 1275 void ConvergingScheduler::SchedBoundary::setLatencyPolicy(CandPolicy &Policy) { 1276 // FIXME: compile time. In all, we visit four queues here one we should only 1277 // need to visit the one that was last popped if we cache the result. 1278 unsigned RemLatency = 0; 1279 for (ReadyQueue::iterator I = Available.begin(), E = Available.end(); 1280 I != E; ++I) { 1281 unsigned L = getUnscheduledLatency(*I); 1282 if (L > RemLatency) 1283 RemLatency = L; 1284 } 1285 for (ReadyQueue::iterator I = Pending.begin(), E = Pending.end(); 1286 I != E; ++I) { 1287 unsigned L = getUnscheduledLatency(*I); 1288 if (L > RemLatency) 1289 RemLatency = L; 1290 } 1291 unsigned CriticalPathLimit = Rem->CriticalPath + SchedModel->getILPWindow(); 1292 if (RemLatency + ExpectedLatency >= CriticalPathLimit 1293 && RemLatency > Rem->getMaxRemainingCount(SchedModel)) { 1294 Policy.ReduceLatency = true; 1295 DEBUG(dbgs() << "Increase ILP: " << Available.getName() << '\n'); 1296 } 1297 } 1298 1299 void ConvergingScheduler::SchedBoundary::releaseNode(SUnit *SU, 1300 unsigned ReadyCycle) { 1301 1302 if (ReadyCycle < MinReadyCycle) 1303 MinReadyCycle = ReadyCycle; 1304 1305 // Check for interlocks first. For the purpose of other heuristics, an 1306 // instruction that cannot issue appears as if it's not in the ReadyQueue. 1307 if (ReadyCycle > CurrCycle || checkHazard(SU)) 1308 Pending.push(SU); 1309 else 1310 Available.push(SU); 1311 1312 // Record this node as an immediate dependent of the scheduled node. 1313 NextSUs.insert(SU); 1314 } 1315 1316 /// Move the boundary of scheduled code by one cycle. 1317 void ConvergingScheduler::SchedBoundary::bumpCycle() { 1318 unsigned Width = SchedModel->getIssueWidth(); 1319 IssueCount = (IssueCount <= Width) ? 0 : IssueCount - Width; 1320 1321 unsigned NextCycle = CurrCycle + 1; 1322 assert(MinReadyCycle < UINT_MAX && "MinReadyCycle uninitialized"); 1323 if (MinReadyCycle > NextCycle) { 1324 IssueCount = 0; 1325 NextCycle = MinReadyCycle; 1326 } 1327 1328 if (!HazardRec->isEnabled()) { 1329 // Bypass HazardRec virtual calls. 1330 CurrCycle = NextCycle; 1331 } 1332 else { 1333 // Bypass getHazardType calls in case of long latency. 1334 for (; CurrCycle != NextCycle; ++CurrCycle) { 1335 if (isTop()) 1336 HazardRec->AdvanceCycle(); 1337 else 1338 HazardRec->RecedeCycle(); 1339 } 1340 } 1341 CheckPending = true; 1342 IsResourceLimited = getCriticalCount() > std::max(ExpectedLatency, CurrCycle); 1343 1344 DEBUG(dbgs() << " *** " << Available.getName() << " cycle " 1345 << CurrCycle << '\n'); 1346 } 1347 1348 /// Add the given processor resource to this scheduled zone. 1349 void ConvergingScheduler::SchedBoundary::countResource(unsigned PIdx, 1350 unsigned Cycles) { 1351 unsigned Factor = SchedModel->getResourceFactor(PIdx); 1352 DEBUG(dbgs() << " " << SchedModel->getProcResource(PIdx)->Name 1353 << " +(" << Cycles << "x" << Factor 1354 << ") / " << SchedModel->getLatencyFactor() << '\n'); 1355 1356 unsigned Count = Factor * Cycles; 1357 ResourceCounts[PIdx] += Count; 1358 assert(Rem->RemainingCounts[PIdx] >= Count && "resource double counted"); 1359 Rem->RemainingCounts[PIdx] -= Count; 1360 1361 // Check if this resource exceeds the current critical resource by a full 1362 // cycle. If so, it becomes the critical resource. 1363 if ((int)(ResourceCounts[PIdx] - ResourceCounts[CritResIdx]) 1364 >= (int)SchedModel->getLatencyFactor()) { 1365 CritResIdx = PIdx; 1366 DEBUG(dbgs() << " *** Critical resource " 1367 << SchedModel->getProcResource(PIdx)->Name << " x" 1368 << ResourceCounts[PIdx] << '\n'); 1369 } 1370 } 1371 1372 /// Move the boundary of scheduled code by one SUnit. 1373 void ConvergingScheduler::SchedBoundary::bumpNode(SUnit *SU) { 1374 // Update the reservation table. 1375 if (HazardRec->isEnabled()) { 1376 if (!isTop() && SU->isCall) { 1377 // Calls are scheduled with their preceding instructions. For bottom-up 1378 // scheduling, clear the pipeline state before emitting. 1379 HazardRec->Reset(); 1380 } 1381 HazardRec->EmitInstruction(SU); 1382 } 1383 // Update resource counts and critical resource. 1384 if (SchedModel->hasInstrSchedModel()) { 1385 const MCSchedClassDesc *SC = DAG->getSchedClass(SU); 1386 Rem->RemainingMicroOps -= SchedModel->getNumMicroOps(SU->getInstr(), SC); 1387 for (TargetSchedModel::ProcResIter 1388 PI = SchedModel->getWriteProcResBegin(SC), 1389 PE = SchedModel->getWriteProcResEnd(SC); PI != PE; ++PI) { 1390 countResource(PI->ProcResourceIdx, PI->Cycles); 1391 } 1392 } 1393 if (isTop()) { 1394 if (SU->getDepth() > ExpectedLatency) 1395 ExpectedLatency = SU->getDepth(); 1396 } 1397 else { 1398 if (SU->getHeight() > ExpectedLatency) 1399 ExpectedLatency = SU->getHeight(); 1400 } 1401 1402 IsResourceLimited = getCriticalCount() > std::max(ExpectedLatency, CurrCycle); 1403 1404 // Check the instruction group dispatch limit. 1405 // TODO: Check if this SU must end a dispatch group. 1406 IssueCount += SchedModel->getNumMicroOps(SU->getInstr()); 1407 1408 // checkHazard prevents scheduling multiple instructions per cycle that exceed 1409 // issue width. However, we commonly reach the maximum. In this case 1410 // opportunistically bump the cycle to avoid uselessly checking everything in 1411 // the readyQ. Furthermore, a single instruction may produce more than one 1412 // cycle's worth of micro-ops. 1413 if (IssueCount >= SchedModel->getIssueWidth()) { 1414 DEBUG(dbgs() << " *** Max instrs at cycle " << CurrCycle << '\n'); 1415 bumpCycle(); 1416 } 1417 } 1418 1419 /// Release pending ready nodes in to the available queue. This makes them 1420 /// visible to heuristics. 1421 void ConvergingScheduler::SchedBoundary::releasePending() { 1422 // If the available queue is empty, it is safe to reset MinReadyCycle. 1423 if (Available.empty()) 1424 MinReadyCycle = UINT_MAX; 1425 1426 // Check to see if any of the pending instructions are ready to issue. If 1427 // so, add them to the available queue. 1428 for (unsigned i = 0, e = Pending.size(); i != e; ++i) { 1429 SUnit *SU = *(Pending.begin()+i); 1430 unsigned ReadyCycle = isTop() ? SU->TopReadyCycle : SU->BotReadyCycle; 1431 1432 if (ReadyCycle < MinReadyCycle) 1433 MinReadyCycle = ReadyCycle; 1434 1435 if (ReadyCycle > CurrCycle) 1436 continue; 1437 1438 if (checkHazard(SU)) 1439 continue; 1440 1441 Available.push(SU); 1442 Pending.remove(Pending.begin()+i); 1443 --i; --e; 1444 } 1445 DEBUG(if (!Pending.empty()) Pending.dump()); 1446 CheckPending = false; 1447 } 1448 1449 /// Remove SU from the ready set for this boundary. 1450 void ConvergingScheduler::SchedBoundary::removeReady(SUnit *SU) { 1451 if (Available.isInQueue(SU)) 1452 Available.remove(Available.find(SU)); 1453 else { 1454 assert(Pending.isInQueue(SU) && "bad ready count"); 1455 Pending.remove(Pending.find(SU)); 1456 } 1457 } 1458 1459 /// If this queue only has one ready candidate, return it. As a side effect, 1460 /// defer any nodes that now hit a hazard, and advance the cycle until at least 1461 /// one node is ready. If multiple instructions are ready, return NULL. 1462 SUnit *ConvergingScheduler::SchedBoundary::pickOnlyChoice() { 1463 if (CheckPending) 1464 releasePending(); 1465 1466 if (IssueCount > 0) { 1467 // Defer any ready instrs that now have a hazard. 1468 for (ReadyQueue::iterator I = Available.begin(); I != Available.end();) { 1469 if (checkHazard(*I)) { 1470 Pending.push(*I); 1471 I = Available.remove(I); 1472 continue; 1473 } 1474 ++I; 1475 } 1476 } 1477 for (unsigned i = 0; Available.empty(); ++i) { 1478 assert(i <= (HazardRec->getMaxLookAhead() + MaxMinLatency) && 1479 "permanent hazard"); (void)i; 1480 bumpCycle(); 1481 releasePending(); 1482 } 1483 if (Available.size() == 1) 1484 return *Available.begin(); 1485 return NULL; 1486 } 1487 1488 /// Record the candidate policy for opposite zones with different critical 1489 /// resources. 1490 /// 1491 /// If the CriticalZone is latency limited, don't force a policy for the 1492 /// candidates here. Instead, setLatencyPolicy sets ReduceLatency if needed. 1493 void ConvergingScheduler::balanceZones( 1494 ConvergingScheduler::SchedBoundary &CriticalZone, 1495 ConvergingScheduler::SchedCandidate &CriticalCand, 1496 ConvergingScheduler::SchedBoundary &OppositeZone, 1497 ConvergingScheduler::SchedCandidate &OppositeCand) { 1498 1499 if (!CriticalZone.IsResourceLimited) 1500 return; 1501 assert(SchedModel->hasInstrSchedModel() && "required schedmodel"); 1502 1503 SchedRemainder *Rem = CriticalZone.Rem; 1504 1505 // If the critical zone is overconsuming a resource relative to the 1506 // remainder, try to reduce it. 1507 unsigned RemainingCritCount = 1508 Rem->RemainingCounts[CriticalZone.CritResIdx]; 1509 if ((int)(Rem->getMaxRemainingCount(SchedModel) - RemainingCritCount) 1510 > (int)SchedModel->getLatencyFactor()) { 1511 CriticalCand.Policy.ReduceResIdx = CriticalZone.CritResIdx; 1512 DEBUG(dbgs() << "Balance " << CriticalZone.Available.getName() << " reduce " 1513 << SchedModel->getProcResource(CriticalZone.CritResIdx)->Name 1514 << '\n'); 1515 } 1516 // If the other zone is underconsuming a resource relative to the full zone, 1517 // try to increase it. 1518 unsigned OppositeCount = 1519 OppositeZone.ResourceCounts[CriticalZone.CritResIdx]; 1520 if ((int)(OppositeZone.ExpectedCount - OppositeCount) 1521 > (int)SchedModel->getLatencyFactor()) { 1522 OppositeCand.Policy.DemandResIdx = CriticalZone.CritResIdx; 1523 DEBUG(dbgs() << "Balance " << OppositeZone.Available.getName() << " demand " 1524 << SchedModel->getProcResource(OppositeZone.CritResIdx)->Name 1525 << '\n'); 1526 } 1527 } 1528 1529 /// Determine if the scheduled zones exceed resource limits or critical path and 1530 /// set each candidate's ReduceHeight policy accordingly. 1531 void ConvergingScheduler::checkResourceLimits( 1532 ConvergingScheduler::SchedCandidate &TopCand, 1533 ConvergingScheduler::SchedCandidate &BotCand) { 1534 1535 // Set ReduceLatency to true if needed. 1536 Bot.setLatencyPolicy(BotCand.Policy); 1537 Top.setLatencyPolicy(TopCand.Policy); 1538 1539 // Handle resource-limited regions. 1540 if (Top.IsResourceLimited && Bot.IsResourceLimited 1541 && Top.CritResIdx == Bot.CritResIdx) { 1542 // If the scheduled critical resource in both zones is no longer the 1543 // critical remaining resource, attempt to reduce resource height both ways. 1544 if (Top.CritResIdx != Rem.CritResIdx) { 1545 TopCand.Policy.ReduceResIdx = Top.CritResIdx; 1546 BotCand.Policy.ReduceResIdx = Bot.CritResIdx; 1547 DEBUG(dbgs() << "Reduce scheduled " 1548 << SchedModel->getProcResource(Top.CritResIdx)->Name << '\n'); 1549 } 1550 return; 1551 } 1552 // Handle latency-limited regions. 1553 if (!Top.IsResourceLimited && !Bot.IsResourceLimited) { 1554 // If the total scheduled expected latency exceeds the region's critical 1555 // path then reduce latency both ways. 1556 // 1557 // Just because a zone is not resource limited does not mean it is latency 1558 // limited. Unbuffered resource, such as max micro-ops may cause CurrCycle 1559 // to exceed expected latency. 1560 if ((Top.ExpectedLatency + Bot.ExpectedLatency >= Rem.CriticalPath) 1561 && (Rem.CriticalPath > Top.CurrCycle + Bot.CurrCycle)) { 1562 TopCand.Policy.ReduceLatency = true; 1563 BotCand.Policy.ReduceLatency = true; 1564 DEBUG(dbgs() << "Reduce scheduled latency " << Top.ExpectedLatency 1565 << " + " << Bot.ExpectedLatency << '\n'); 1566 } 1567 return; 1568 } 1569 // The critical resource is different in each zone, so request balancing. 1570 1571 // Compute the cost of each zone. 1572 Top.ExpectedCount = std::max(Top.ExpectedLatency, Top.CurrCycle); 1573 Top.ExpectedCount = std::max( 1574 Top.getCriticalCount(), 1575 Top.ExpectedCount * SchedModel->getLatencyFactor()); 1576 Bot.ExpectedCount = std::max(Bot.ExpectedLatency, Bot.CurrCycle); 1577 Bot.ExpectedCount = std::max( 1578 Bot.getCriticalCount(), 1579 Bot.ExpectedCount * SchedModel->getLatencyFactor()); 1580 1581 balanceZones(Top, TopCand, Bot, BotCand); 1582 balanceZones(Bot, BotCand, Top, TopCand); 1583 } 1584 1585 void ConvergingScheduler::SchedCandidate:: 1586 initResourceDelta(const ScheduleDAGMI *DAG, 1587 const TargetSchedModel *SchedModel) { 1588 if (!Policy.ReduceResIdx && !Policy.DemandResIdx) 1589 return; 1590 1591 const MCSchedClassDesc *SC = DAG->getSchedClass(SU); 1592 for (TargetSchedModel::ProcResIter 1593 PI = SchedModel->getWriteProcResBegin(SC), 1594 PE = SchedModel->getWriteProcResEnd(SC); PI != PE; ++PI) { 1595 if (PI->ProcResourceIdx == Policy.ReduceResIdx) 1596 ResDelta.CritResources += PI->Cycles; 1597 if (PI->ProcResourceIdx == Policy.DemandResIdx) 1598 ResDelta.DemandedResources += PI->Cycles; 1599 } 1600 } 1601 1602 /// Return true if this heuristic determines order. 1603 static bool tryLess(unsigned TryVal, unsigned CandVal, 1604 ConvergingScheduler::SchedCandidate &TryCand, 1605 ConvergingScheduler::SchedCandidate &Cand, 1606 ConvergingScheduler::CandReason Reason) { 1607 if (TryVal < CandVal) { 1608 TryCand.Reason = Reason; 1609 return true; 1610 } 1611 if (TryVal > CandVal) { 1612 if (Cand.Reason > Reason) 1613 Cand.Reason = Reason; 1614 return true; 1615 } 1616 return false; 1617 } 1618 1619 static bool tryGreater(unsigned TryVal, unsigned CandVal, 1620 ConvergingScheduler::SchedCandidate &TryCand, 1621 ConvergingScheduler::SchedCandidate &Cand, 1622 ConvergingScheduler::CandReason Reason) { 1623 if (TryVal > CandVal) { 1624 TryCand.Reason = Reason; 1625 return true; 1626 } 1627 if (TryVal < CandVal) { 1628 if (Cand.Reason > Reason) 1629 Cand.Reason = Reason; 1630 return true; 1631 } 1632 return false; 1633 } 1634 1635 static unsigned getWeakLeft(const SUnit *SU, bool isTop) { 1636 return (isTop) ? SU->WeakPredsLeft : SU->WeakSuccsLeft; 1637 } 1638 1639 /// Apply a set of heursitics to a new candidate. Heuristics are currently 1640 /// hierarchical. This may be more efficient than a graduated cost model because 1641 /// we don't need to evaluate all aspects of the model for each node in the 1642 /// queue. But it's really done to make the heuristics easier to debug and 1643 /// statistically analyze. 1644 /// 1645 /// \param Cand provides the policy and current best candidate. 1646 /// \param TryCand refers to the next SUnit candidate, otherwise uninitialized. 1647 /// \param Zone describes the scheduled zone that we are extending. 1648 /// \param RPTracker describes reg pressure within the scheduled zone. 1649 /// \param TempTracker is a scratch pressure tracker to reuse in queries. 1650 void ConvergingScheduler::tryCandidate(SchedCandidate &Cand, 1651 SchedCandidate &TryCand, 1652 SchedBoundary &Zone, 1653 const RegPressureTracker &RPTracker, 1654 RegPressureTracker &TempTracker) { 1655 1656 // Always initialize TryCand's RPDelta. 1657 TempTracker.getMaxPressureDelta(TryCand.SU->getInstr(), TryCand.RPDelta, 1658 DAG->getRegionCriticalPSets(), 1659 DAG->getRegPressure().MaxSetPressure); 1660 1661 // Initialize the candidate if needed. 1662 if (!Cand.isValid()) { 1663 TryCand.Reason = NodeOrder; 1664 return; 1665 } 1666 // Avoid exceeding the target's limit. 1667 if (tryLess(TryCand.RPDelta.Excess.UnitIncrease, 1668 Cand.RPDelta.Excess.UnitIncrease, TryCand, Cand, SingleExcess)) 1669 return; 1670 if (Cand.Reason == SingleExcess) 1671 Cand.Reason = MultiPressure; 1672 1673 // Avoid increasing the max critical pressure in the scheduled region. 1674 if (tryLess(TryCand.RPDelta.CriticalMax.UnitIncrease, 1675 Cand.RPDelta.CriticalMax.UnitIncrease, 1676 TryCand, Cand, SingleCritical)) 1677 return; 1678 if (Cand.Reason == SingleCritical) 1679 Cand.Reason = MultiPressure; 1680 1681 // Keep clustered nodes together to encourage downstream peephole 1682 // optimizations which may reduce resource requirements. 1683 // 1684 // This is a best effort to set things up for a post-RA pass. Optimizations 1685 // like generating loads of multiple registers should ideally be done within 1686 // the scheduler pass by combining the loads during DAG postprocessing. 1687 const SUnit *NextClusterSU = 1688 Zone.isTop() ? DAG->getNextClusterSucc() : DAG->getNextClusterPred(); 1689 if (tryGreater(TryCand.SU == NextClusterSU, Cand.SU == NextClusterSU, 1690 TryCand, Cand, Cluster)) 1691 return; 1692 // Currently, weak edges are for clustering, so we hard-code that reason. 1693 // However, deferring the current TryCand will not change Cand's reason. 1694 CandReason OrigReason = Cand.Reason; 1695 if (tryLess(getWeakLeft(TryCand.SU, Zone.isTop()), 1696 getWeakLeft(Cand.SU, Zone.isTop()), 1697 TryCand, Cand, Cluster)) { 1698 Cand.Reason = OrigReason; 1699 return; 1700 } 1701 // Avoid critical resource consumption and balance the schedule. 1702 TryCand.initResourceDelta(DAG, SchedModel); 1703 if (tryLess(TryCand.ResDelta.CritResources, Cand.ResDelta.CritResources, 1704 TryCand, Cand, ResourceReduce)) 1705 return; 1706 if (tryGreater(TryCand.ResDelta.DemandedResources, 1707 Cand.ResDelta.DemandedResources, 1708 TryCand, Cand, ResourceDemand)) 1709 return; 1710 1711 // Avoid serializing long latency dependence chains. 1712 if (Cand.Policy.ReduceLatency) { 1713 if (Zone.isTop()) { 1714 if (Cand.SU->getDepth() * SchedModel->getLatencyFactor() 1715 > Zone.ExpectedCount) { 1716 if (tryLess(TryCand.SU->getDepth(), Cand.SU->getDepth(), 1717 TryCand, Cand, TopDepthReduce)) 1718 return; 1719 } 1720 if (tryGreater(TryCand.SU->getHeight(), Cand.SU->getHeight(), 1721 TryCand, Cand, TopPathReduce)) 1722 return; 1723 } 1724 else { 1725 if (Cand.SU->getHeight() * SchedModel->getLatencyFactor() 1726 > Zone.ExpectedCount) { 1727 if (tryLess(TryCand.SU->getHeight(), Cand.SU->getHeight(), 1728 TryCand, Cand, BotHeightReduce)) 1729 return; 1730 } 1731 if (tryGreater(TryCand.SU->getDepth(), Cand.SU->getDepth(), 1732 TryCand, Cand, BotPathReduce)) 1733 return; 1734 } 1735 } 1736 1737 // Avoid increasing the max pressure of the entire region. 1738 if (tryLess(TryCand.RPDelta.CurrentMax.UnitIncrease, 1739 Cand.RPDelta.CurrentMax.UnitIncrease, TryCand, Cand, SingleMax)) 1740 return; 1741 if (Cand.Reason == SingleMax) 1742 Cand.Reason = MultiPressure; 1743 1744 // Prefer immediate defs/users of the last scheduled instruction. This is a 1745 // nice pressure avoidance strategy that also conserves the processor's 1746 // register renaming resources and keeps the machine code readable. 1747 if (tryGreater(Zone.NextSUs.count(TryCand.SU), Zone.NextSUs.count(Cand.SU), 1748 TryCand, Cand, NextDefUse)) 1749 return; 1750 1751 // Fall through to original instruction order. 1752 if ((Zone.isTop() && TryCand.SU->NodeNum < Cand.SU->NodeNum) 1753 || (!Zone.isTop() && TryCand.SU->NodeNum > Cand.SU->NodeNum)) { 1754 TryCand.Reason = NodeOrder; 1755 } 1756 } 1757 1758 /// pickNodeFromQueue helper that returns true if the LHS reg pressure effect is 1759 /// more desirable than RHS from scheduling standpoint. 1760 static bool compareRPDelta(const RegPressureDelta &LHS, 1761 const RegPressureDelta &RHS) { 1762 // Compare each component of pressure in decreasing order of importance 1763 // without checking if any are valid. Invalid PressureElements are assumed to 1764 // have UnitIncrease==0, so are neutral. 1765 1766 // Avoid increasing the max critical pressure in the scheduled region. 1767 if (LHS.Excess.UnitIncrease != RHS.Excess.UnitIncrease) { 1768 DEBUG(dbgs() << "RP excess top - bot: " 1769 << (LHS.Excess.UnitIncrease - RHS.Excess.UnitIncrease) << '\n'); 1770 return LHS.Excess.UnitIncrease < RHS.Excess.UnitIncrease; 1771 } 1772 // Avoid increasing the max critical pressure in the scheduled region. 1773 if (LHS.CriticalMax.UnitIncrease != RHS.CriticalMax.UnitIncrease) { 1774 DEBUG(dbgs() << "RP critical top - bot: " 1775 << (LHS.CriticalMax.UnitIncrease - RHS.CriticalMax.UnitIncrease) 1776 << '\n'); 1777 return LHS.CriticalMax.UnitIncrease < RHS.CriticalMax.UnitIncrease; 1778 } 1779 // Avoid increasing the max pressure of the entire region. 1780 if (LHS.CurrentMax.UnitIncrease != RHS.CurrentMax.UnitIncrease) { 1781 DEBUG(dbgs() << "RP current top - bot: " 1782 << (LHS.CurrentMax.UnitIncrease - RHS.CurrentMax.UnitIncrease) 1783 << '\n'); 1784 return LHS.CurrentMax.UnitIncrease < RHS.CurrentMax.UnitIncrease; 1785 } 1786 return false; 1787 } 1788 1789 #ifndef NDEBUG 1790 const char *ConvergingScheduler::getReasonStr( 1791 ConvergingScheduler::CandReason Reason) { 1792 switch (Reason) { 1793 case NoCand: return "NOCAND "; 1794 case SingleExcess: return "REG-EXCESS"; 1795 case SingleCritical: return "REG-CRIT "; 1796 case Cluster: return "CLUSTER "; 1797 case SingleMax: return "REG-MAX "; 1798 case MultiPressure: return "REG-MULTI "; 1799 case ResourceReduce: return "RES-REDUCE"; 1800 case ResourceDemand: return "RES-DEMAND"; 1801 case TopDepthReduce: return "TOP-DEPTH "; 1802 case TopPathReduce: return "TOP-PATH "; 1803 case BotHeightReduce:return "BOT-HEIGHT"; 1804 case BotPathReduce: return "BOT-PATH "; 1805 case NextDefUse: return "DEF-USE "; 1806 case NodeOrder: return "ORDER "; 1807 }; 1808 llvm_unreachable("Unknown reason!"); 1809 } 1810 1811 void ConvergingScheduler::traceCandidate(const SchedCandidate &Cand, 1812 const SchedBoundary &Zone) { 1813 const char *Label = getReasonStr(Cand.Reason); 1814 PressureElement P; 1815 unsigned ResIdx = 0; 1816 unsigned Latency = 0; 1817 switch (Cand.Reason) { 1818 default: 1819 break; 1820 case SingleExcess: 1821 P = Cand.RPDelta.Excess; 1822 break; 1823 case SingleCritical: 1824 P = Cand.RPDelta.CriticalMax; 1825 break; 1826 case SingleMax: 1827 P = Cand.RPDelta.CurrentMax; 1828 break; 1829 case ResourceReduce: 1830 ResIdx = Cand.Policy.ReduceResIdx; 1831 break; 1832 case ResourceDemand: 1833 ResIdx = Cand.Policy.DemandResIdx; 1834 break; 1835 case TopDepthReduce: 1836 Latency = Cand.SU->getDepth(); 1837 break; 1838 case TopPathReduce: 1839 Latency = Cand.SU->getHeight(); 1840 break; 1841 case BotHeightReduce: 1842 Latency = Cand.SU->getHeight(); 1843 break; 1844 case BotPathReduce: 1845 Latency = Cand.SU->getDepth(); 1846 break; 1847 } 1848 dbgs() << Label << " " << Zone.Available.getName() << " "; 1849 if (P.isValid()) 1850 dbgs() << TRI->getRegPressureSetName(P.PSetID) << ":" << P.UnitIncrease 1851 << " "; 1852 else 1853 dbgs() << " "; 1854 if (ResIdx) 1855 dbgs() << SchedModel->getProcResource(ResIdx)->Name << " "; 1856 else 1857 dbgs() << " "; 1858 if (Latency) 1859 dbgs() << Latency << " cycles "; 1860 else 1861 dbgs() << " "; 1862 Cand.SU->dump(DAG); 1863 } 1864 #endif 1865 1866 /// Pick the best candidate from the top queue. 1867 /// 1868 /// TODO: getMaxPressureDelta results can be mostly cached for each SUnit during 1869 /// DAG building. To adjust for the current scheduling location we need to 1870 /// maintain the number of vreg uses remaining to be top-scheduled. 1871 void ConvergingScheduler::pickNodeFromQueue(SchedBoundary &Zone, 1872 const RegPressureTracker &RPTracker, 1873 SchedCandidate &Cand) { 1874 ReadyQueue &Q = Zone.Available; 1875 1876 DEBUG(Q.dump()); 1877 1878 // getMaxPressureDelta temporarily modifies the tracker. 1879 RegPressureTracker &TempTracker = const_cast<RegPressureTracker&>(RPTracker); 1880 1881 for (ReadyQueue::iterator I = Q.begin(), E = Q.end(); I != E; ++I) { 1882 1883 SchedCandidate TryCand(Cand.Policy); 1884 TryCand.SU = *I; 1885 tryCandidate(Cand, TryCand, Zone, RPTracker, TempTracker); 1886 if (TryCand.Reason != NoCand) { 1887 // Initialize resource delta if needed in case future heuristics query it. 1888 if (TryCand.ResDelta == SchedResourceDelta()) 1889 TryCand.initResourceDelta(DAG, SchedModel); 1890 Cand.setBest(TryCand); 1891 DEBUG(traceCandidate(Cand, Zone)); 1892 } 1893 } 1894 } 1895 1896 static void tracePick(const ConvergingScheduler::SchedCandidate &Cand, 1897 bool IsTop) { 1898 DEBUG(dbgs() << "Pick " << (IsTop ? "top" : "bot") 1899 << " SU(" << Cand.SU->NodeNum << ") " 1900 << ConvergingScheduler::getReasonStr(Cand.Reason) << '\n'); 1901 } 1902 1903 /// Pick the best candidate node from either the top or bottom queue. 1904 SUnit *ConvergingScheduler::pickNodeBidirectional(bool &IsTopNode) { 1905 // Schedule as far as possible in the direction of no choice. This is most 1906 // efficient, but also provides the best heuristics for CriticalPSets. 1907 if (SUnit *SU = Bot.pickOnlyChoice()) { 1908 IsTopNode = false; 1909 return SU; 1910 } 1911 if (SUnit *SU = Top.pickOnlyChoice()) { 1912 IsTopNode = true; 1913 return SU; 1914 } 1915 CandPolicy NoPolicy; 1916 SchedCandidate BotCand(NoPolicy); 1917 SchedCandidate TopCand(NoPolicy); 1918 checkResourceLimits(TopCand, BotCand); 1919 1920 // Prefer bottom scheduling when heuristics are silent. 1921 pickNodeFromQueue(Bot, DAG->getBotRPTracker(), BotCand); 1922 assert(BotCand.Reason != NoCand && "failed to find the first candidate"); 1923 1924 // If either Q has a single candidate that provides the least increase in 1925 // Excess pressure, we can immediately schedule from that Q. 1926 // 1927 // RegionCriticalPSets summarizes the pressure within the scheduled region and 1928 // affects picking from either Q. If scheduling in one direction must 1929 // increase pressure for one of the excess PSets, then schedule in that 1930 // direction first to provide more freedom in the other direction. 1931 if (BotCand.Reason == SingleExcess || BotCand.Reason == SingleCritical) { 1932 IsTopNode = false; 1933 tracePick(BotCand, IsTopNode); 1934 return BotCand.SU; 1935 } 1936 // Check if the top Q has a better candidate. 1937 pickNodeFromQueue(Top, DAG->getTopRPTracker(), TopCand); 1938 assert(TopCand.Reason != NoCand && "failed to find the first candidate"); 1939 1940 // If either Q has a single candidate that minimizes pressure above the 1941 // original region's pressure pick it. 1942 if (TopCand.Reason <= SingleMax || BotCand.Reason <= SingleMax) { 1943 if (TopCand.Reason < BotCand.Reason) { 1944 IsTopNode = true; 1945 tracePick(TopCand, IsTopNode); 1946 return TopCand.SU; 1947 } 1948 IsTopNode = false; 1949 tracePick(BotCand, IsTopNode); 1950 return BotCand.SU; 1951 } 1952 // Check for a salient pressure difference and pick the best from either side. 1953 if (compareRPDelta(TopCand.RPDelta, BotCand.RPDelta)) { 1954 IsTopNode = true; 1955 tracePick(TopCand, IsTopNode); 1956 return TopCand.SU; 1957 } 1958 // Otherwise prefer the bottom candidate, in node order if all else failed. 1959 if (TopCand.Reason < BotCand.Reason) { 1960 IsTopNode = true; 1961 tracePick(TopCand, IsTopNode); 1962 return TopCand.SU; 1963 } 1964 IsTopNode = false; 1965 tracePick(BotCand, IsTopNode); 1966 return BotCand.SU; 1967 } 1968 1969 /// Pick the best node to balance the schedule. Implements MachineSchedStrategy. 1970 SUnit *ConvergingScheduler::pickNode(bool &IsTopNode) { 1971 if (DAG->top() == DAG->bottom()) { 1972 assert(Top.Available.empty() && Top.Pending.empty() && 1973 Bot.Available.empty() && Bot.Pending.empty() && "ReadyQ garbage"); 1974 return NULL; 1975 } 1976 SUnit *SU; 1977 do { 1978 if (ForceTopDown) { 1979 SU = Top.pickOnlyChoice(); 1980 if (!SU) { 1981 CandPolicy NoPolicy; 1982 SchedCandidate TopCand(NoPolicy); 1983 pickNodeFromQueue(Top, DAG->getTopRPTracker(), TopCand); 1984 assert(TopCand.Reason != NoCand && "failed to find the first candidate"); 1985 SU = TopCand.SU; 1986 } 1987 IsTopNode = true; 1988 } 1989 else if (ForceBottomUp) { 1990 SU = Bot.pickOnlyChoice(); 1991 if (!SU) { 1992 CandPolicy NoPolicy; 1993 SchedCandidate BotCand(NoPolicy); 1994 pickNodeFromQueue(Bot, DAG->getBotRPTracker(), BotCand); 1995 assert(BotCand.Reason != NoCand && "failed to find the first candidate"); 1996 SU = BotCand.SU; 1997 } 1998 IsTopNode = false; 1999 } 2000 else { 2001 SU = pickNodeBidirectional(IsTopNode); 2002 } 2003 } while (SU->isScheduled); 2004 2005 if (SU->isTopReady()) 2006 Top.removeReady(SU); 2007 if (SU->isBottomReady()) 2008 Bot.removeReady(SU); 2009 2010 DEBUG(dbgs() << "*** " << (IsTopNode ? "Top" : "Bottom") 2011 << " Scheduling Instruction in cycle " 2012 << (IsTopNode ? Top.CurrCycle : Bot.CurrCycle) << '\n'; 2013 SU->dump(DAG)); 2014 return SU; 2015 } 2016 2017 /// Update the scheduler's state after scheduling a node. This is the same node 2018 /// that was just returned by pickNode(). However, ScheduleDAGMI needs to update 2019 /// it's state based on the current cycle before MachineSchedStrategy does. 2020 void ConvergingScheduler::schedNode(SUnit *SU, bool IsTopNode) { 2021 if (IsTopNode) { 2022 SU->TopReadyCycle = Top.CurrCycle; 2023 Top.bumpNode(SU); 2024 } 2025 else { 2026 SU->BotReadyCycle = Bot.CurrCycle; 2027 Bot.bumpNode(SU); 2028 } 2029 } 2030 2031 /// Create the standard converging machine scheduler. This will be used as the 2032 /// default scheduler if the target does not set a default. 2033 static ScheduleDAGInstrs *createConvergingSched(MachineSchedContext *C) { 2034 assert((!ForceTopDown || !ForceBottomUp) && 2035 "-misched-topdown incompatible with -misched-bottomup"); 2036 ScheduleDAGMI *DAG = new ScheduleDAGMI(C, new ConvergingScheduler()); 2037 // Register DAG post-processors. 2038 if (EnableLoadCluster) 2039 DAG->addMutation(new LoadClusterMutation(DAG->TII, DAG->TRI)); 2040 if (EnableMacroFusion) 2041 DAG->addMutation(new MacroFusion(DAG->TII)); 2042 return DAG; 2043 } 2044 static MachineSchedRegistry 2045 ConvergingSchedRegistry("converge", "Standard converging scheduler.", 2046 createConvergingSched); 2047 2048 //===----------------------------------------------------------------------===// 2049 // ILP Scheduler. Currently for experimental analysis of heuristics. 2050 //===----------------------------------------------------------------------===// 2051 2052 namespace { 2053 /// \brief Order nodes by the ILP metric. 2054 struct ILPOrder { 2055 SchedDFSResult *DFSResult; 2056 BitVector *ScheduledTrees; 2057 bool MaximizeILP; 2058 2059 ILPOrder(SchedDFSResult *dfs, BitVector *schedtrees, bool MaxILP) 2060 : DFSResult(dfs), ScheduledTrees(schedtrees), MaximizeILP(MaxILP) {} 2061 2062 /// \brief Apply a less-than relation on node priority. 2063 /// 2064 /// (Return true if A comes after B in the Q.) 2065 bool operator()(const SUnit *A, const SUnit *B) const { 2066 unsigned SchedTreeA = DFSResult->getSubtreeID(A); 2067 unsigned SchedTreeB = DFSResult->getSubtreeID(B); 2068 if (SchedTreeA != SchedTreeB) { 2069 // Unscheduled trees have lower priority. 2070 if (ScheduledTrees->test(SchedTreeA) != ScheduledTrees->test(SchedTreeB)) 2071 return ScheduledTrees->test(SchedTreeB); 2072 2073 // Trees with shallower connections have have lower priority. 2074 if (DFSResult->getSubtreeLevel(SchedTreeA) 2075 != DFSResult->getSubtreeLevel(SchedTreeB)) { 2076 return DFSResult->getSubtreeLevel(SchedTreeA) 2077 < DFSResult->getSubtreeLevel(SchedTreeB); 2078 } 2079 } 2080 if (MaximizeILP) 2081 return DFSResult->getILP(A) < DFSResult->getILP(B); 2082 else 2083 return DFSResult->getILP(A) > DFSResult->getILP(B); 2084 } 2085 }; 2086 2087 /// \brief Schedule based on the ILP metric. 2088 class ILPScheduler : public MachineSchedStrategy { 2089 /// In case all subtrees are eventually connected to a common root through 2090 /// data dependence (e.g. reduction), place an upper limit on their size. 2091 /// 2092 /// FIXME: A subtree limit is generally good, but in the situation commented 2093 /// above, where multiple similar subtrees feed a common root, we should 2094 /// only split at a point where the resulting subtrees will be balanced. 2095 /// (a motivating test case must be found). 2096 static const unsigned SubtreeLimit = 16; 2097 2098 SchedDFSResult DFSResult; 2099 BitVector ScheduledTrees; 2100 ILPOrder Cmp; 2101 2102 std::vector<SUnit*> ReadyQ; 2103 public: 2104 ILPScheduler(bool MaximizeILP) 2105 : DFSResult(/*BottomUp=*/true, SubtreeLimit), 2106 Cmp(&DFSResult, &ScheduledTrees, MaximizeILP) {} 2107 2108 virtual void initialize(ScheduleDAGMI *DAG) { 2109 ReadyQ.clear(); 2110 DFSResult.clear(); 2111 DFSResult.resize(DAG->SUnits.size()); 2112 ScheduledTrees.clear(); 2113 } 2114 2115 virtual void registerRoots() { 2116 DFSResult.compute(ReadyQ); 2117 ScheduledTrees.resize(DFSResult.getNumSubtrees()); 2118 // Restore the heap in ReadyQ with the updated DFS results. 2119 std::make_heap(ReadyQ.begin(), ReadyQ.end(), Cmp); 2120 } 2121 2122 /// Implement MachineSchedStrategy interface. 2123 /// ----------------------------------------- 2124 2125 /// Callback to select the highest priority node from the ready Q. 2126 virtual SUnit *pickNode(bool &IsTopNode) { 2127 if (ReadyQ.empty()) return NULL; 2128 pop_heap(ReadyQ.begin(), ReadyQ.end(), Cmp); 2129 SUnit *SU = ReadyQ.back(); 2130 ReadyQ.pop_back(); 2131 IsTopNode = false; 2132 DEBUG(dbgs() << "*** Scheduling " << "SU(" << SU->NodeNum << "): " 2133 << *SU->getInstr() 2134 << " ILP: " << DFSResult.getILP(SU) 2135 << " Tree: " << DFSResult.getSubtreeID(SU) << " @" 2136 << DFSResult.getSubtreeLevel(DFSResult.getSubtreeID(SU))<< '\n'); 2137 return SU; 2138 } 2139 2140 /// Callback after a node is scheduled. Mark a newly scheduled tree, notify 2141 /// DFSResults, and resort the priority Q. 2142 virtual void schedNode(SUnit *SU, bool IsTopNode) { 2143 assert(!IsTopNode && "SchedDFSResult needs bottom-up"); 2144 if (!ScheduledTrees.test(DFSResult.getSubtreeID(SU))) { 2145 ScheduledTrees.set(DFSResult.getSubtreeID(SU)); 2146 DFSResult.scheduleTree(DFSResult.getSubtreeID(SU)); 2147 std::make_heap(ReadyQ.begin(), ReadyQ.end(), Cmp); 2148 } 2149 } 2150 2151 virtual void releaseTopNode(SUnit *) { /*only called for top roots*/ } 2152 2153 virtual void releaseBottomNode(SUnit *SU) { 2154 ReadyQ.push_back(SU); 2155 std::push_heap(ReadyQ.begin(), ReadyQ.end(), Cmp); 2156 } 2157 }; 2158 } // namespace 2159 2160 static ScheduleDAGInstrs *createILPMaxScheduler(MachineSchedContext *C) { 2161 return new ScheduleDAGMI(C, new ILPScheduler(true)); 2162 } 2163 static ScheduleDAGInstrs *createILPMinScheduler(MachineSchedContext *C) { 2164 return new ScheduleDAGMI(C, new ILPScheduler(false)); 2165 } 2166 static MachineSchedRegistry ILPMaxRegistry( 2167 "ilpmax", "Schedule bottom-up for max ILP", createILPMaxScheduler); 2168 static MachineSchedRegistry ILPMinRegistry( 2169 "ilpmin", "Schedule bottom-up for min ILP", createILPMinScheduler); 2170 2171 //===----------------------------------------------------------------------===// 2172 // Machine Instruction Shuffler for Correctness Testing 2173 //===----------------------------------------------------------------------===// 2174 2175 #ifndef NDEBUG 2176 namespace { 2177 /// Apply a less-than relation on the node order, which corresponds to the 2178 /// instruction order prior to scheduling. IsReverse implements greater-than. 2179 template<bool IsReverse> 2180 struct SUnitOrder { 2181 bool operator()(SUnit *A, SUnit *B) const { 2182 if (IsReverse) 2183 return A->NodeNum > B->NodeNum; 2184 else 2185 return A->NodeNum < B->NodeNum; 2186 } 2187 }; 2188 2189 /// Reorder instructions as much as possible. 2190 class InstructionShuffler : public MachineSchedStrategy { 2191 bool IsAlternating; 2192 bool IsTopDown; 2193 2194 // Using a less-than relation (SUnitOrder<false>) for the TopQ priority 2195 // gives nodes with a higher number higher priority causing the latest 2196 // instructions to be scheduled first. 2197 PriorityQueue<SUnit*, std::vector<SUnit*>, SUnitOrder<false> > 2198 TopQ; 2199 // When scheduling bottom-up, use greater-than as the queue priority. 2200 PriorityQueue<SUnit*, std::vector<SUnit*>, SUnitOrder<true> > 2201 BottomQ; 2202 public: 2203 InstructionShuffler(bool alternate, bool topdown) 2204 : IsAlternating(alternate), IsTopDown(topdown) {} 2205 2206 virtual void initialize(ScheduleDAGMI *) { 2207 TopQ.clear(); 2208 BottomQ.clear(); 2209 } 2210 2211 /// Implement MachineSchedStrategy interface. 2212 /// ----------------------------------------- 2213 2214 virtual SUnit *pickNode(bool &IsTopNode) { 2215 SUnit *SU; 2216 if (IsTopDown) { 2217 do { 2218 if (TopQ.empty()) return NULL; 2219 SU = TopQ.top(); 2220 TopQ.pop(); 2221 } while (SU->isScheduled); 2222 IsTopNode = true; 2223 } 2224 else { 2225 do { 2226 if (BottomQ.empty()) return NULL; 2227 SU = BottomQ.top(); 2228 BottomQ.pop(); 2229 } while (SU->isScheduled); 2230 IsTopNode = false; 2231 } 2232 if (IsAlternating) 2233 IsTopDown = !IsTopDown; 2234 return SU; 2235 } 2236 2237 virtual void schedNode(SUnit *SU, bool IsTopNode) {} 2238 2239 virtual void releaseTopNode(SUnit *SU) { 2240 TopQ.push(SU); 2241 } 2242 virtual void releaseBottomNode(SUnit *SU) { 2243 BottomQ.push(SU); 2244 } 2245 }; 2246 } // namespace 2247 2248 static ScheduleDAGInstrs *createInstructionShuffler(MachineSchedContext *C) { 2249 bool Alternate = !ForceTopDown && !ForceBottomUp; 2250 bool TopDown = !ForceBottomUp; 2251 assert((TopDown || !ForceTopDown) && 2252 "-misched-topdown incompatible with -misched-bottomup"); 2253 return new ScheduleDAGMI(C, new InstructionShuffler(Alternate, TopDown)); 2254 } 2255 static MachineSchedRegistry ShufflerRegistry( 2256 "shuffle", "Shuffle machine instructions alternating directions", 2257 createInstructionShuffler); 2258 #endif // !NDEBUG 2259