1 //===-- R600MachineScheduler.cpp - R600 Scheduler Interface -*- C++ -*-----===//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
8 //
9 /// \file
10 /// R600 Machine Scheduler interface
11 //
12 //===----------------------------------------------------------------------===//
13 
14 #include "R600MachineScheduler.h"
15 #include "MCTargetDesc/R600MCTargetDesc.h"
16 #include "R600Subtarget.h"
17 
18 using namespace llvm;
19 
20 #define DEBUG_TYPE "machine-scheduler"
21 
22 void R600SchedStrategy::initialize(ScheduleDAGMI *dag) {
23   assert(dag->hasVRegLiveness() && "R600SchedStrategy needs vreg liveness");
24   DAG = static_cast<ScheduleDAGMILive*>(dag);
25   const R600Subtarget &ST = DAG->MF.getSubtarget<R600Subtarget>();
26   TII = static_cast<const R600InstrInfo*>(DAG->TII);
27   TRI = static_cast<const R600RegisterInfo*>(DAG->TRI);
28   VLIW5 = !ST.hasCaymanISA();
29   MRI = &DAG->MRI;
30   CurInstKind = IDOther;
31   CurEmitted = 0;
32   OccupiedSlotsMask = 31;
33   InstKindLimit[IDAlu] = TII->getMaxAlusPerClause();
34   InstKindLimit[IDOther] = 32;
35   InstKindLimit[IDFetch] = ST.getTexVTXClauseSize();
36   AluInstCount = 0;
37   FetchInstCount = 0;
38 }
39 
40 void R600SchedStrategy::MoveUnits(std::vector<SUnit *> &QSrc,
41                                   std::vector<SUnit *> &QDst)
42 {
43   llvm::append_range(QDst, QSrc);
44   QSrc.clear();
45 }
46 
47 static unsigned getWFCountLimitedByGPR(unsigned GPRCount) {
48   assert (GPRCount && "GPRCount cannot be 0");
49   return 248 / GPRCount;
50 }
51 
52 SUnit* R600SchedStrategy::pickNode(bool &IsTopNode) {
53   SUnit *SU = nullptr;
54   NextInstKind = IDOther;
55 
56   IsTopNode = false;
57 
58   // check if we might want to switch current clause type
59   bool AllowSwitchToAlu = (CurEmitted >= InstKindLimit[CurInstKind]) ||
60       (Available[CurInstKind].empty());
61   bool AllowSwitchFromAlu = (CurEmitted >= InstKindLimit[CurInstKind]) &&
62       (!Available[IDFetch].empty() || !Available[IDOther].empty());
63 
64   if (CurInstKind == IDAlu && !Available[IDFetch].empty()) {
65     // We use the heuristic provided by AMD Accelerated Parallel Processing
66     // OpenCL Programming Guide :
67     // The approx. number of WF that allows TEX inst to hide ALU inst is :
68     // 500 (cycles for TEX) / (AluFetchRatio * 8 (cycles for ALU))
69     float ALUFetchRationEstimate =
70         (AluInstCount + AvailablesAluCount() + Pending[IDAlu].size()) /
71         (FetchInstCount + Available[IDFetch].size());
72     if (ALUFetchRationEstimate == 0) {
73       AllowSwitchFromAlu = true;
74     } else {
75       unsigned NeededWF = 62.5f / ALUFetchRationEstimate;
76       LLVM_DEBUG(dbgs() << NeededWF << " approx. Wavefronts Required\n");
77       // We assume the local GPR requirements to be "dominated" by the requirement
78       // of the TEX clause (which consumes 128 bits regs) ; ALU inst before and
79       // after TEX are indeed likely to consume or generate values from/for the
80       // TEX clause.
81       // Available[IDFetch].size() * 2 : GPRs required in the Fetch clause
82       // We assume that fetch instructions are either TnXYZW = TEX TnXYZW (need
83       // one GPR) or TmXYZW = TnXYZW (need 2 GPR).
84       // (TODO : use RegisterPressure)
85       // If we are going too use too many GPR, we flush Fetch instruction to lower
86       // register pressure on 128 bits regs.
87       unsigned NearRegisterRequirement = 2 * Available[IDFetch].size();
88       if (NeededWF > getWFCountLimitedByGPR(NearRegisterRequirement))
89         AllowSwitchFromAlu = true;
90     }
91   }
92 
93   if (!SU && ((AllowSwitchToAlu && CurInstKind != IDAlu) ||
94       (!AllowSwitchFromAlu && CurInstKind == IDAlu))) {
95     // try to pick ALU
96     SU = pickAlu();
97     if (!SU && !PhysicalRegCopy.empty()) {
98       SU = PhysicalRegCopy.front();
99       PhysicalRegCopy.erase(PhysicalRegCopy.begin());
100     }
101     if (SU) {
102       if (CurEmitted >= InstKindLimit[IDAlu])
103         CurEmitted = 0;
104       NextInstKind = IDAlu;
105     }
106   }
107 
108   if (!SU) {
109     // try to pick FETCH
110     SU = pickOther(IDFetch);
111     if (SU)
112       NextInstKind = IDFetch;
113   }
114 
115   // try to pick other
116   if (!SU) {
117     SU = pickOther(IDOther);
118     if (SU)
119       NextInstKind = IDOther;
120   }
121 
122   LLVM_DEBUG(if (SU) {
123     dbgs() << " ** Pick node **\n";
124     DAG->dumpNode(*SU);
125   } else {
126     dbgs() << "NO NODE \n";
127     for (const SUnit &S : DAG->SUnits)
128       if (!S.isScheduled)
129         DAG->dumpNode(S);
130   });
131 
132   return SU;
133 }
134 
135 void R600SchedStrategy::schedNode(SUnit *SU, bool IsTopNode) {
136   if (NextInstKind != CurInstKind) {
137     LLVM_DEBUG(dbgs() << "Instruction Type Switch\n");
138     if (NextInstKind != IDAlu)
139       OccupiedSlotsMask |= 31;
140     CurEmitted = 0;
141     CurInstKind = NextInstKind;
142   }
143 
144   if (CurInstKind == IDAlu) {
145     AluInstCount ++;
146     switch (getAluKind(SU)) {
147     case AluT_XYZW:
148       CurEmitted += 4;
149       break;
150     case AluDiscarded:
151       break;
152     default: {
153       ++CurEmitted;
154       for (MachineInstr::mop_iterator It = SU->getInstr()->operands_begin(),
155           E = SU->getInstr()->operands_end(); It != E; ++It) {
156         MachineOperand &MO = *It;
157         if (MO.isReg() && MO.getReg() == R600::ALU_LITERAL_X)
158           ++CurEmitted;
159       }
160     }
161     }
162   } else {
163     ++CurEmitted;
164   }
165 
166   LLVM_DEBUG(dbgs() << CurEmitted << " Instructions Emitted in this clause\n");
167 
168   if (CurInstKind != IDFetch) {
169     MoveUnits(Pending[IDFetch], Available[IDFetch]);
170   } else
171     FetchInstCount++;
172 }
173 
174 static bool
175 isPhysicalRegCopy(MachineInstr *MI) {
176   if (MI->getOpcode() != R600::COPY)
177     return false;
178 
179   return !MI->getOperand(1).getReg().isVirtual();
180 }
181 
182 void R600SchedStrategy::releaseTopNode(SUnit *SU) {
183   LLVM_DEBUG(dbgs() << "Top Releasing "; DAG->dumpNode(*SU));
184 }
185 
186 void R600SchedStrategy::releaseBottomNode(SUnit *SU) {
187   LLVM_DEBUG(dbgs() << "Bottom Releasing "; DAG->dumpNode(*SU));
188   if (isPhysicalRegCopy(SU->getInstr())) {
189     PhysicalRegCopy.push_back(SU);
190     return;
191   }
192 
193   int IK = getInstKind(SU);
194 
195   // There is no export clause, we can schedule one as soon as its ready
196   if (IK == IDOther)
197     Available[IDOther].push_back(SU);
198   else
199     Pending[IK].push_back(SU);
200 
201 }
202 
203 bool R600SchedStrategy::regBelongsToClass(Register Reg,
204                                           const TargetRegisterClass *RC) const {
205   if (!Reg.isVirtual()) {
206     return RC->contains(Reg);
207   } else {
208     return MRI->getRegClass(Reg) == RC;
209   }
210 }
211 
212 R600SchedStrategy::AluKind R600SchedStrategy::getAluKind(SUnit *SU) const {
213   MachineInstr *MI = SU->getInstr();
214 
215   if (TII->isTransOnly(*MI))
216     return AluTrans;
217 
218   switch (MI->getOpcode()) {
219   case R600::PRED_X:
220     return AluPredX;
221   case R600::INTERP_PAIR_XY:
222   case R600::INTERP_PAIR_ZW:
223   case R600::INTERP_VEC_LOAD:
224   case R600::DOT_4:
225     return AluT_XYZW;
226   case R600::COPY:
227     if (MI->getOperand(1).isUndef()) {
228       // MI will become a KILL, don't considers it in scheduling
229       return AluDiscarded;
230     }
231     break;
232   default:
233     break;
234   }
235 
236   // Does the instruction take a whole IG ?
237   // XXX: Is it possible to add a helper function in R600InstrInfo that can
238   // be used here and in R600PacketizerList::isSoloInstruction() ?
239   if(TII->isVector(*MI) ||
240      TII->isCubeOp(MI->getOpcode()) ||
241      TII->isReductionOp(MI->getOpcode()) ||
242      MI->getOpcode() == R600::GROUP_BARRIER) {
243     return AluT_XYZW;
244   }
245 
246   if (TII->isLDSInstr(MI->getOpcode())) {
247     return AluT_X;
248   }
249 
250   // Is the result already assigned to a channel ?
251   unsigned DestSubReg = MI->getOperand(0).getSubReg();
252   switch (DestSubReg) {
253   case R600::sub0:
254     return AluT_X;
255   case R600::sub1:
256     return AluT_Y;
257   case R600::sub2:
258     return AluT_Z;
259   case R600::sub3:
260     return AluT_W;
261   default:
262     break;
263   }
264 
265   // Is the result already member of a X/Y/Z/W class ?
266   Register DestReg = MI->getOperand(0).getReg();
267   if (regBelongsToClass(DestReg, &R600::R600_TReg32_XRegClass) ||
268       regBelongsToClass(DestReg, &R600::R600_AddrRegClass))
269     return AluT_X;
270   if (regBelongsToClass(DestReg, &R600::R600_TReg32_YRegClass))
271     return AluT_Y;
272   if (regBelongsToClass(DestReg, &R600::R600_TReg32_ZRegClass))
273     return AluT_Z;
274   if (regBelongsToClass(DestReg, &R600::R600_TReg32_WRegClass))
275     return AluT_W;
276   if (regBelongsToClass(DestReg, &R600::R600_Reg128RegClass))
277     return AluT_XYZW;
278 
279   // LDS src registers cannot be used in the Trans slot.
280   if (TII->readsLDSSrcReg(*MI))
281     return AluT_XYZW;
282 
283   return AluAny;
284 }
285 
286 int R600SchedStrategy::getInstKind(SUnit* SU) {
287   int Opcode = SU->getInstr()->getOpcode();
288 
289   if (TII->usesTextureCache(Opcode) || TII->usesVertexCache(Opcode))
290     return IDFetch;
291 
292   if (TII->isALUInstr(Opcode)) {
293     return IDAlu;
294   }
295 
296   switch (Opcode) {
297   case R600::PRED_X:
298   case R600::COPY:
299   case R600::CONST_COPY:
300   case R600::INTERP_PAIR_XY:
301   case R600::INTERP_PAIR_ZW:
302   case R600::INTERP_VEC_LOAD:
303   case R600::DOT_4:
304     return IDAlu;
305   default:
306     return IDOther;
307   }
308 }
309 
310 SUnit *R600SchedStrategy::PopInst(std::vector<SUnit *> &Q, bool AnyALU) {
311   if (Q.empty())
312     return nullptr;
313   for (std::vector<SUnit *>::reverse_iterator It = Q.rbegin(), E = Q.rend();
314       It != E; ++It) {
315     SUnit *SU = *It;
316     InstructionsGroupCandidate.push_back(SU->getInstr());
317     if (TII->fitsConstReadLimitations(InstructionsGroupCandidate) &&
318         (!AnyALU || !TII->isVectorOnly(*SU->getInstr()))) {
319       InstructionsGroupCandidate.pop_back();
320       Q.erase((It + 1).base());
321       return SU;
322     } else {
323       InstructionsGroupCandidate.pop_back();
324     }
325   }
326   return nullptr;
327 }
328 
329 void R600SchedStrategy::LoadAlu() {
330   std::vector<SUnit *> &QSrc = Pending[IDAlu];
331   for (SUnit *SU : QSrc) {
332     AluKind AK = getAluKind(SU);
333     AvailableAlus[AK].push_back(SU);
334   }
335   QSrc.clear();
336 }
337 
338 void R600SchedStrategy::PrepareNextSlot() {
339   LLVM_DEBUG(dbgs() << "New Slot\n");
340   assert(OccupiedSlotsMask && "Slot wasn't filled");
341   OccupiedSlotsMask = 0;
342   //  if (HwGen == AMDGPUSubtarget::NORTHERN_ISLANDS)
343   //    OccupiedSlotsMask |= 16;
344   InstructionsGroupCandidate.clear();
345   LoadAlu();
346 }
347 
348 void R600SchedStrategy::AssignSlot(MachineInstr* MI, unsigned Slot) {
349   int DstIndex = TII->getOperandIdx(MI->getOpcode(), R600::OpName::dst);
350   if (DstIndex == -1) {
351     return;
352   }
353   Register DestReg = MI->getOperand(DstIndex).getReg();
354   // PressureRegister crashes if an operand is def and used in the same inst
355   // and we try to constraint its regclass
356   for (MachineInstr::mop_iterator It = MI->operands_begin(),
357       E = MI->operands_end(); It != E; ++It) {
358     MachineOperand &MO = *It;
359     if (MO.isReg() && !MO.isDef() &&
360         MO.getReg() == DestReg)
361       return;
362   }
363   // Constrains the regclass of DestReg to assign it to Slot
364   switch (Slot) {
365   case 0:
366     MRI->constrainRegClass(DestReg, &R600::R600_TReg32_XRegClass);
367     break;
368   case 1:
369     MRI->constrainRegClass(DestReg, &R600::R600_TReg32_YRegClass);
370     break;
371   case 2:
372     MRI->constrainRegClass(DestReg, &R600::R600_TReg32_ZRegClass);
373     break;
374   case 3:
375     MRI->constrainRegClass(DestReg, &R600::R600_TReg32_WRegClass);
376     break;
377   }
378 }
379 
380 SUnit *R600SchedStrategy::AttemptFillSlot(unsigned Slot, bool AnyAlu) {
381   static const AluKind IndexToID[] = {AluT_X, AluT_Y, AluT_Z, AluT_W};
382   SUnit *SlotedSU = PopInst(AvailableAlus[IndexToID[Slot]], AnyAlu);
383   if (SlotedSU)
384     return SlotedSU;
385   SUnit *UnslotedSU = PopInst(AvailableAlus[AluAny], AnyAlu);
386   if (UnslotedSU)
387     AssignSlot(UnslotedSU->getInstr(), Slot);
388   return UnslotedSU;
389 }
390 
391 unsigned R600SchedStrategy::AvailablesAluCount() const {
392   return AvailableAlus[AluAny].size() + AvailableAlus[AluT_XYZW].size() +
393       AvailableAlus[AluT_X].size() + AvailableAlus[AluT_Y].size() +
394       AvailableAlus[AluT_Z].size() + AvailableAlus[AluT_W].size() +
395       AvailableAlus[AluTrans].size() + AvailableAlus[AluDiscarded].size() +
396       AvailableAlus[AluPredX].size();
397 }
398 
399 SUnit* R600SchedStrategy::pickAlu() {
400   while (AvailablesAluCount() || !Pending[IDAlu].empty()) {
401     if (!OccupiedSlotsMask) {
402       // Bottom up scheduling : predX must comes first
403       if (!AvailableAlus[AluPredX].empty()) {
404         OccupiedSlotsMask |= 31;
405         return PopInst(AvailableAlus[AluPredX], false);
406       }
407       // Flush physical reg copies (RA will discard them)
408       if (!AvailableAlus[AluDiscarded].empty()) {
409         OccupiedSlotsMask |= 31;
410         return PopInst(AvailableAlus[AluDiscarded], false);
411       }
412       // If there is a T_XYZW alu available, use it
413       if (!AvailableAlus[AluT_XYZW].empty()) {
414         OccupiedSlotsMask |= 15;
415         return PopInst(AvailableAlus[AluT_XYZW], false);
416       }
417     }
418     bool TransSlotOccupied = OccupiedSlotsMask & 16;
419     if (!TransSlotOccupied && VLIW5) {
420       if (!AvailableAlus[AluTrans].empty()) {
421         OccupiedSlotsMask |= 16;
422         return PopInst(AvailableAlus[AluTrans], false);
423       }
424       SUnit *SU = AttemptFillSlot(3, true);
425       if (SU) {
426         OccupiedSlotsMask |= 16;
427         return SU;
428       }
429     }
430     for (int Chan = 3; Chan > -1; --Chan) {
431       bool isOccupied = OccupiedSlotsMask & (1 << Chan);
432       if (!isOccupied) {
433         SUnit *SU = AttemptFillSlot(Chan, false);
434         if (SU) {
435           OccupiedSlotsMask |= (1 << Chan);
436           InstructionsGroupCandidate.push_back(SU->getInstr());
437           return SU;
438         }
439       }
440     }
441     PrepareNextSlot();
442   }
443   return nullptr;
444 }
445 
446 SUnit* R600SchedStrategy::pickOther(int QID) {
447   SUnit *SU = nullptr;
448   std::vector<SUnit *> &AQ = Available[QID];
449 
450   if (AQ.empty()) {
451     MoveUnits(Pending[QID], AQ);
452   }
453   if (!AQ.empty()) {
454     SU = AQ.back();
455     AQ.pop_back();
456   }
457   return SU;
458 }
459