LLVM 22.0.0git
VPlanTransforms.cpp
Go to the documentation of this file.
1//===-- VPlanTransforms.cpp - Utility VPlan to VPlan transforms -----------===//
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/// This file implements a set of utility VPlan to VPlan transformations.
11///
12//===----------------------------------------------------------------------===//
13
14#include "VPlanTransforms.h"
15#include "VPRecipeBuilder.h"
16#include "VPlan.h"
17#include "VPlanAnalysis.h"
18#include "VPlanCFG.h"
19#include "VPlanDominatorTree.h"
20#include "VPlanHelpers.h"
21#include "VPlanPatternMatch.h"
22#include "VPlanUtils.h"
23#include "VPlanVerifier.h"
24#include "llvm/ADT/APInt.h"
26#include "llvm/ADT/STLExtras.h"
27#include "llvm/ADT/SetVector.h"
28#include "llvm/ADT/TypeSwitch.h"
34#include "llvm/IR/Intrinsics.h"
35#include "llvm/IR/MDBuilder.h"
39
40using namespace llvm;
41using namespace VPlanPatternMatch;
42
44 "enable-wide-lane-mask", cl::init(false), cl::Hidden,
45 cl::desc("Enable use of wide get active lane mask instructions"));
46
48 VPlanPtr &Plan,
50 GetIntOrFpInductionDescriptor,
51 const TargetLibraryInfo &TLI) {
52
54 Plan->getVectorLoopRegion());
56 // Skip blocks outside region
57 if (!VPBB->getParent())
58 break;
59 VPRecipeBase *Term = VPBB->getTerminator();
60 auto EndIter = Term ? Term->getIterator() : VPBB->end();
61 // Introduce each ingredient into VPlan.
62 for (VPRecipeBase &Ingredient :
63 make_early_inc_range(make_range(VPBB->begin(), EndIter))) {
64
65 VPValue *VPV = Ingredient.getVPSingleValue();
66 if (!VPV->getUnderlyingValue())
67 continue;
68
70
71 VPRecipeBase *NewRecipe = nullptr;
72 if (auto *PhiR = dyn_cast<VPPhi>(&Ingredient)) {
73 auto *Phi = cast<PHINode>(PhiR->getUnderlyingValue());
74 const auto *II = GetIntOrFpInductionDescriptor(Phi);
75 if (!II) {
76 NewRecipe = new VPWidenPHIRecipe(Phi, nullptr, PhiR->getDebugLoc());
77 for (VPValue *Op : PhiR->operands())
78 NewRecipe->addOperand(Op);
79 } else {
80 VPValue *Start = Plan->getOrAddLiveIn(II->getStartValue());
81 VPValue *Step =
83 NewRecipe = new VPWidenIntOrFpInductionRecipe(
84 Phi, Start, Step, &Plan->getVF(), *II, Ingredient.getDebugLoc());
85 }
86 } else {
87 assert(isa<VPInstruction>(&Ingredient) &&
88 "only VPInstructions expected here");
89 assert(!isa<PHINode>(Inst) && "phis should be handled above");
90 // Create VPWidenMemoryRecipe for loads and stores.
91 if (LoadInst *Load = dyn_cast<LoadInst>(Inst)) {
92 NewRecipe = new VPWidenLoadRecipe(
93 *Load, Ingredient.getOperand(0), nullptr /*Mask*/,
94 false /*Consecutive*/, false /*Reverse*/, VPIRMetadata(*Load),
95 Ingredient.getDebugLoc());
96 } else if (StoreInst *Store = dyn_cast<StoreInst>(Inst)) {
97 NewRecipe = new VPWidenStoreRecipe(
98 *Store, Ingredient.getOperand(1), Ingredient.getOperand(0),
99 nullptr /*Mask*/, false /*Consecutive*/, false /*Reverse*/,
100 VPIRMetadata(*Store), Ingredient.getDebugLoc());
102 NewRecipe = new VPWidenGEPRecipe(GEP, Ingredient.operands());
103 } else if (CallInst *CI = dyn_cast<CallInst>(Inst)) {
104 Intrinsic::ID VectorID = getVectorIntrinsicIDForCall(CI, &TLI);
105 if (VectorID == Intrinsic::not_intrinsic)
106 return false;
107 NewRecipe = new VPWidenIntrinsicRecipe(
108 *CI, getVectorIntrinsicIDForCall(CI, &TLI),
109 {Ingredient.op_begin(), Ingredient.op_end() - 1}, CI->getType(),
110 CI->getDebugLoc());
111 } else if (SelectInst *SI = dyn_cast<SelectInst>(Inst)) {
112 NewRecipe = new VPWidenSelectRecipe(*SI, Ingredient.operands());
113 } else if (auto *CI = dyn_cast<CastInst>(Inst)) {
114 NewRecipe = new VPWidenCastRecipe(
115 CI->getOpcode(), Ingredient.getOperand(0), CI->getType(), *CI);
116 } else {
117 NewRecipe = new VPWidenRecipe(*Inst, Ingredient.operands());
118 }
119 }
120
121 NewRecipe->insertBefore(&Ingredient);
122 if (NewRecipe->getNumDefinedValues() == 1)
123 VPV->replaceAllUsesWith(NewRecipe->getVPSingleValue());
124 else
125 assert(NewRecipe->getNumDefinedValues() == 0 &&
126 "Only recpies with zero or one defined values expected");
127 Ingredient.eraseFromParent();
128 }
129 }
130 return true;
131}
132
133static bool sinkScalarOperands(VPlan &Plan) {
134 auto Iter = vp_depth_first_deep(Plan.getEntry());
135 bool Changed = false;
136 // First, collect the operands of all recipes in replicate blocks as seeds for
137 // sinking.
140 VPBasicBlock *EntryVPBB = VPR->getEntryBasicBlock();
141 if (!VPR->isReplicator() || EntryVPBB->getSuccessors().size() != 2)
142 continue;
143 VPBasicBlock *VPBB = dyn_cast<VPBasicBlock>(EntryVPBB->getSuccessors()[0]);
144 if (!VPBB || VPBB->getSingleSuccessor() != VPR->getExitingBasicBlock())
145 continue;
146 for (auto &Recipe : *VPBB) {
147 for (VPValue *Op : Recipe.operands())
148 if (auto *Def =
149 dyn_cast_or_null<VPSingleDefRecipe>(Op->getDefiningRecipe()))
150 WorkList.insert({VPBB, Def});
151 }
152 }
153
154 bool ScalarVFOnly = Plan.hasScalarVFOnly();
155 // Try to sink each replicate or scalar IV steps recipe in the worklist.
156 for (unsigned I = 0; I != WorkList.size(); ++I) {
157 VPBasicBlock *SinkTo;
158 VPSingleDefRecipe *SinkCandidate;
159 std::tie(SinkTo, SinkCandidate) = WorkList[I];
160 if (SinkCandidate->getParent() == SinkTo ||
161 SinkCandidate->mayHaveSideEffects() ||
162 SinkCandidate->mayReadOrWriteMemory())
163 continue;
164 if (auto *RepR = dyn_cast<VPReplicateRecipe>(SinkCandidate)) {
165 if (!ScalarVFOnly && RepR->isSingleScalar())
166 continue;
167 } else if (!isa<VPScalarIVStepsRecipe>(SinkCandidate))
168 continue;
169
170 bool NeedsDuplicating = false;
171 // All recipe users of the sink candidate must be in the same block SinkTo
172 // or all users outside of SinkTo must be uniform-after-vectorization (
173 // i.e., only first lane is used) . In the latter case, we need to duplicate
174 // SinkCandidate.
175 auto CanSinkWithUser = [SinkTo, &NeedsDuplicating,
176 SinkCandidate](VPUser *U) {
177 auto *UI = cast<VPRecipeBase>(U);
178 if (UI->getParent() == SinkTo)
179 return true;
180 NeedsDuplicating = UI->onlyFirstLaneUsed(SinkCandidate);
181 // We only know how to duplicate VPReplicateRecipes and
182 // VPScalarIVStepsRecipes for now.
183 return NeedsDuplicating &&
185 };
186 if (!all_of(SinkCandidate->users(), CanSinkWithUser))
187 continue;
188
189 if (NeedsDuplicating) {
190 if (ScalarVFOnly)
191 continue;
192 VPSingleDefRecipe *Clone;
193 if (auto *SinkCandidateRepR =
194 dyn_cast<VPReplicateRecipe>(SinkCandidate)) {
195 // TODO: Handle converting to uniform recipes as separate transform,
196 // then cloning should be sufficient here.
197 Instruction *I = SinkCandidate->getUnderlyingInstr();
198 Clone = new VPReplicateRecipe(I, SinkCandidate->operands(), true,
199 nullptr /*Mask*/, *SinkCandidateRepR);
200 // TODO: add ".cloned" suffix to name of Clone's VPValue.
201 } else {
202 Clone = SinkCandidate->clone();
203 }
204
205 Clone->insertBefore(SinkCandidate);
206 SinkCandidate->replaceUsesWithIf(Clone, [SinkTo](VPUser &U, unsigned) {
207 return cast<VPRecipeBase>(&U)->getParent() != SinkTo;
208 });
209 }
210 SinkCandidate->moveBefore(*SinkTo, SinkTo->getFirstNonPhi());
211 for (VPValue *Op : SinkCandidate->operands())
212 if (auto *Def =
213 dyn_cast_or_null<VPSingleDefRecipe>(Op->getDefiningRecipe()))
214 WorkList.insert({SinkTo, Def});
215 Changed = true;
216 }
217 return Changed;
218}
219
220/// If \p R is a region with a VPBranchOnMaskRecipe in the entry block, return
221/// the mask.
223 auto *EntryBB = dyn_cast<VPBasicBlock>(R->getEntry());
224 if (!EntryBB || EntryBB->size() != 1 ||
225 !isa<VPBranchOnMaskRecipe>(EntryBB->begin()))
226 return nullptr;
227
228 return cast<VPBranchOnMaskRecipe>(&*EntryBB->begin())->getOperand(0);
229}
230
231/// If \p R is a triangle region, return the 'then' block of the triangle.
233 auto *EntryBB = cast<VPBasicBlock>(R->getEntry());
234 if (EntryBB->getNumSuccessors() != 2)
235 return nullptr;
236
237 auto *Succ0 = dyn_cast<VPBasicBlock>(EntryBB->getSuccessors()[0]);
238 auto *Succ1 = dyn_cast<VPBasicBlock>(EntryBB->getSuccessors()[1]);
239 if (!Succ0 || !Succ1)
240 return nullptr;
241
242 if (Succ0->getNumSuccessors() + Succ1->getNumSuccessors() != 1)
243 return nullptr;
244 if (Succ0->getSingleSuccessor() == Succ1)
245 return Succ0;
246 if (Succ1->getSingleSuccessor() == Succ0)
247 return Succ1;
248 return nullptr;
249}
250
251// Merge replicate regions in their successor region, if a replicate region
252// is connected to a successor replicate region with the same predicate by a
253// single, empty VPBasicBlock.
255 SmallPtrSet<VPRegionBlock *, 4> TransformedRegions;
256
257 // Collect replicate regions followed by an empty block, followed by another
258 // replicate region with matching masks to process front. This is to avoid
259 // iterator invalidation issues while merging regions.
262 vp_depth_first_deep(Plan.getEntry()))) {
263 if (!Region1->isReplicator())
264 continue;
265 auto *MiddleBasicBlock =
266 dyn_cast_or_null<VPBasicBlock>(Region1->getSingleSuccessor());
267 if (!MiddleBasicBlock || !MiddleBasicBlock->empty())
268 continue;
269
270 auto *Region2 =
271 dyn_cast_or_null<VPRegionBlock>(MiddleBasicBlock->getSingleSuccessor());
272 if (!Region2 || !Region2->isReplicator())
273 continue;
274
275 VPValue *Mask1 = getPredicatedMask(Region1);
276 VPValue *Mask2 = getPredicatedMask(Region2);
277 if (!Mask1 || Mask1 != Mask2)
278 continue;
279
280 assert(Mask1 && Mask2 && "both region must have conditions");
281 WorkList.push_back(Region1);
282 }
283
284 // Move recipes from Region1 to its successor region, if both are triangles.
285 for (VPRegionBlock *Region1 : WorkList) {
286 if (TransformedRegions.contains(Region1))
287 continue;
288 auto *MiddleBasicBlock = cast<VPBasicBlock>(Region1->getSingleSuccessor());
289 auto *Region2 = cast<VPRegionBlock>(MiddleBasicBlock->getSingleSuccessor());
290
291 VPBasicBlock *Then1 = getPredicatedThenBlock(Region1);
292 VPBasicBlock *Then2 = getPredicatedThenBlock(Region2);
293 if (!Then1 || !Then2)
294 continue;
295
296 // Note: No fusion-preventing memory dependencies are expected in either
297 // region. Such dependencies should be rejected during earlier dependence
298 // checks, which guarantee accesses can be re-ordered for vectorization.
299 //
300 // Move recipes to the successor region.
301 for (VPRecipeBase &ToMove : make_early_inc_range(reverse(*Then1)))
302 ToMove.moveBefore(*Then2, Then2->getFirstNonPhi());
303
304 auto *Merge1 = cast<VPBasicBlock>(Then1->getSingleSuccessor());
305 auto *Merge2 = cast<VPBasicBlock>(Then2->getSingleSuccessor());
306
307 // Move VPPredInstPHIRecipes from the merge block to the successor region's
308 // merge block. Update all users inside the successor region to use the
309 // original values.
310 for (VPRecipeBase &Phi1ToMove : make_early_inc_range(reverse(*Merge1))) {
311 VPValue *PredInst1 =
312 cast<VPPredInstPHIRecipe>(&Phi1ToMove)->getOperand(0);
313 VPValue *Phi1ToMoveV = Phi1ToMove.getVPSingleValue();
314 Phi1ToMoveV->replaceUsesWithIf(PredInst1, [Then2](VPUser &U, unsigned) {
315 return cast<VPRecipeBase>(&U)->getParent() == Then2;
316 });
317
318 // Remove phi recipes that are unused after merging the regions.
319 if (Phi1ToMove.getVPSingleValue()->getNumUsers() == 0) {
320 Phi1ToMove.eraseFromParent();
321 continue;
322 }
323 Phi1ToMove.moveBefore(*Merge2, Merge2->begin());
324 }
325
326 // Remove the dead recipes in Region1's entry block.
327 for (VPRecipeBase &R :
328 make_early_inc_range(reverse(*Region1->getEntryBasicBlock())))
329 R.eraseFromParent();
330
331 // Finally, remove the first region.
332 for (VPBlockBase *Pred : make_early_inc_range(Region1->getPredecessors())) {
333 VPBlockUtils::disconnectBlocks(Pred, Region1);
334 VPBlockUtils::connectBlocks(Pred, MiddleBasicBlock);
335 }
336 VPBlockUtils::disconnectBlocks(Region1, MiddleBasicBlock);
337 TransformedRegions.insert(Region1);
338 }
339
340 return !TransformedRegions.empty();
341}
342
344 VPlan &Plan) {
345 Instruction *Instr = PredRecipe->getUnderlyingInstr();
346 // Build the triangular if-then region.
347 std::string RegionName = (Twine("pred.") + Instr->getOpcodeName()).str();
348 assert(Instr->getParent() && "Predicated instruction not in any basic block");
349 auto *BlockInMask = PredRecipe->getMask();
350 auto *MaskDef = BlockInMask->getDefiningRecipe();
351 auto *BOMRecipe = new VPBranchOnMaskRecipe(
352 BlockInMask, MaskDef ? MaskDef->getDebugLoc() : DebugLoc::getUnknown());
353 auto *Entry =
354 Plan.createVPBasicBlock(Twine(RegionName) + ".entry", BOMRecipe);
355
356 // Replace predicated replicate recipe with a replicate recipe without a
357 // mask but in the replicate region.
358 auto *RecipeWithoutMask = new VPReplicateRecipe(
359 PredRecipe->getUnderlyingInstr(),
360 make_range(PredRecipe->op_begin(), std::prev(PredRecipe->op_end())),
361 PredRecipe->isSingleScalar(), nullptr /*Mask*/, *PredRecipe);
362 auto *Pred =
363 Plan.createVPBasicBlock(Twine(RegionName) + ".if", RecipeWithoutMask);
364
365 VPPredInstPHIRecipe *PHIRecipe = nullptr;
366 if (PredRecipe->getNumUsers() != 0) {
367 PHIRecipe = new VPPredInstPHIRecipe(RecipeWithoutMask,
368 RecipeWithoutMask->getDebugLoc());
369 PredRecipe->replaceAllUsesWith(PHIRecipe);
370 PHIRecipe->setOperand(0, RecipeWithoutMask);
371 }
372 PredRecipe->eraseFromParent();
373 auto *Exiting =
374 Plan.createVPBasicBlock(Twine(RegionName) + ".continue", PHIRecipe);
376 Plan.createVPRegionBlock(Entry, Exiting, RegionName, true);
377
378 // Note: first set Entry as region entry and then connect successors starting
379 // from it in order, to propagate the "parent" of each VPBasicBlock.
380 VPBlockUtils::insertTwoBlocksAfter(Pred, Exiting, Entry);
381 VPBlockUtils::connectBlocks(Pred, Exiting);
382
383 return Region;
384}
385
386static void addReplicateRegions(VPlan &Plan) {
389 vp_depth_first_deep(Plan.getEntry()))) {
390 for (VPRecipeBase &R : *VPBB)
391 if (auto *RepR = dyn_cast<VPReplicateRecipe>(&R)) {
392 if (RepR->isPredicated())
393 WorkList.push_back(RepR);
394 }
395 }
396
397 unsigned BBNum = 0;
398 for (VPReplicateRecipe *RepR : WorkList) {
399 VPBasicBlock *CurrentBlock = RepR->getParent();
400 VPBasicBlock *SplitBlock = CurrentBlock->splitAt(RepR->getIterator());
401
402 BasicBlock *OrigBB = RepR->getUnderlyingInstr()->getParent();
403 SplitBlock->setName(
404 OrigBB->hasName() ? OrigBB->getName() + "." + Twine(BBNum++) : "");
405 // Record predicated instructions for above packing optimizations.
407 Region->setParent(CurrentBlock->getParent());
409
410 VPRegionBlock *ParentRegion = Region->getParent();
411 if (ParentRegion && ParentRegion->getExiting() == CurrentBlock)
412 ParentRegion->setExiting(SplitBlock);
413 }
414}
415
416/// Remove redundant VPBasicBlocks by merging them into their predecessor if
417/// the predecessor has a single successor.
421 vp_depth_first_deep(Plan.getEntry()))) {
422 // Don't fold the blocks in the skeleton of the Plan into their single
423 // predecessors for now.
424 // TODO: Remove restriction once more of the skeleton is modeled in VPlan.
425 if (!VPBB->getParent())
426 continue;
427 auto *PredVPBB =
428 dyn_cast_or_null<VPBasicBlock>(VPBB->getSinglePredecessor());
429 if (!PredVPBB || PredVPBB->getNumSuccessors() != 1 ||
430 isa<VPIRBasicBlock>(PredVPBB))
431 continue;
432 WorkList.push_back(VPBB);
433 }
434
435 for (VPBasicBlock *VPBB : WorkList) {
436 VPBasicBlock *PredVPBB = cast<VPBasicBlock>(VPBB->getSinglePredecessor());
437 for (VPRecipeBase &R : make_early_inc_range(*VPBB))
438 R.moveBefore(*PredVPBB, PredVPBB->end());
439 VPBlockUtils::disconnectBlocks(PredVPBB, VPBB);
440 auto *ParentRegion = VPBB->getParent();
441 if (ParentRegion && ParentRegion->getExiting() == VPBB)
442 ParentRegion->setExiting(PredVPBB);
443 for (auto *Succ : to_vector(VPBB->successors())) {
445 VPBlockUtils::connectBlocks(PredVPBB, Succ);
446 }
447 // VPBB is now dead and will be cleaned up when the plan gets destroyed.
448 }
449 return !WorkList.empty();
450}
451
453 // Convert masked VPReplicateRecipes to if-then region blocks.
455
456 bool ShouldSimplify = true;
457 while (ShouldSimplify) {
458 ShouldSimplify = sinkScalarOperands(Plan);
459 ShouldSimplify |= mergeReplicateRegionsIntoSuccessors(Plan);
460 ShouldSimplify |= mergeBlocksIntoPredecessors(Plan);
461 }
462}
463
464/// Remove redundant casts of inductions.
465///
466/// Such redundant casts are casts of induction variables that can be ignored,
467/// because we already proved that the casted phi is equal to the uncasted phi
468/// in the vectorized loop. There is no need to vectorize the cast - the same
469/// value can be used for both the phi and casts in the vector loop.
471 for (auto &Phi : Plan.getVectorLoopRegion()->getEntryBasicBlock()->phis()) {
473 if (!IV || IV->getTruncInst())
474 continue;
475
476 // A sequence of IR Casts has potentially been recorded for IV, which
477 // *must be bypassed* when the IV is vectorized, because the vectorized IV
478 // will produce the desired casted value. This sequence forms a def-use
479 // chain and is provided in reverse order, ending with the cast that uses
480 // the IV phi. Search for the recipe of the last cast in the chain and
481 // replace it with the original IV. Note that only the final cast is
482 // expected to have users outside the cast-chain and the dead casts left
483 // over will be cleaned up later.
484 auto &Casts = IV->getInductionDescriptor().getCastInsts();
485 VPValue *FindMyCast = IV;
486 for (Instruction *IRCast : reverse(Casts)) {
487 VPSingleDefRecipe *FoundUserCast = nullptr;
488 for (auto *U : FindMyCast->users()) {
489 auto *UserCast = dyn_cast<VPSingleDefRecipe>(U);
490 if (UserCast && UserCast->getUnderlyingValue() == IRCast) {
491 FoundUserCast = UserCast;
492 break;
493 }
494 }
495 FindMyCast = FoundUserCast;
496 }
497 FindMyCast->replaceAllUsesWith(IV);
498 }
499}
500
501/// Try to replace VPWidenCanonicalIVRecipes with a widened canonical IV
502/// recipe, if it exists.
504 VPCanonicalIVPHIRecipe *CanonicalIV = Plan.getCanonicalIV();
505 VPWidenCanonicalIVRecipe *WidenNewIV = nullptr;
506 for (VPUser *U : CanonicalIV->users()) {
508 if (WidenNewIV)
509 break;
510 }
511
512 if (!WidenNewIV)
513 return;
514
516 for (VPRecipeBase &Phi : HeaderVPBB->phis()) {
517 auto *WidenOriginalIV = dyn_cast<VPWidenIntOrFpInductionRecipe>(&Phi);
518
519 if (!WidenOriginalIV || !WidenOriginalIV->isCanonical())
520 continue;
521
522 // Replace WidenNewIV with WidenOriginalIV if WidenOriginalIV provides
523 // everything WidenNewIV's users need. That is, WidenOriginalIV will
524 // generate a vector phi or all users of WidenNewIV demand the first lane
525 // only.
526 if (!vputils::onlyScalarValuesUsed(WidenOriginalIV) ||
527 vputils::onlyFirstLaneUsed(WidenNewIV)) {
528 WidenNewIV->replaceAllUsesWith(WidenOriginalIV);
529 WidenNewIV->eraseFromParent();
530 return;
531 }
532 }
533}
534
535/// Returns true if \p R is dead and can be removed.
536static bool isDeadRecipe(VPRecipeBase &R) {
537 // Do remove conditional assume instructions as their conditions may be
538 // flattened.
539 auto *RepR = dyn_cast<VPReplicateRecipe>(&R);
540 bool IsConditionalAssume = RepR && RepR->isPredicated() &&
542 if (IsConditionalAssume)
543 return true;
544
545 if (R.mayHaveSideEffects())
546 return false;
547
548 // Recipe is dead if no user keeps the recipe alive.
549 return all_of(R.definedValues(),
550 [](VPValue *V) { return V->getNumUsers() == 0; });
551}
552
555 vp_post_order_deep(Plan.getEntry()))) {
556 // The recipes in the block are processed in reverse order, to catch chains
557 // of dead recipes.
558 for (VPRecipeBase &R : make_early_inc_range(reverse(*VPBB))) {
559 if (isDeadRecipe(R)) {
560 R.eraseFromParent();
561 continue;
562 }
563
564 // Check if R is a dead VPPhi <-> update cycle and remove it.
565 auto *PhiR = dyn_cast<VPPhi>(&R);
566 if (!PhiR || PhiR->getNumOperands() != 2 || PhiR->getNumUsers() != 1)
567 continue;
568 VPValue *Incoming = PhiR->getOperand(1);
569 if (*PhiR->user_begin() != Incoming->getDefiningRecipe() ||
570 Incoming->getNumUsers() != 1)
571 continue;
572 PhiR->replaceAllUsesWith(PhiR->getOperand(0));
573 PhiR->eraseFromParent();
574 Incoming->getDefiningRecipe()->eraseFromParent();
575 }
576 }
577}
578
581 Instruction::BinaryOps InductionOpcode,
582 FPMathOperator *FPBinOp, Instruction *TruncI,
583 VPValue *StartV, VPValue *Step, DebugLoc DL,
584 VPBuilder &Builder) {
586 VPCanonicalIVPHIRecipe *CanonicalIV = Plan.getCanonicalIV();
587 VPSingleDefRecipe *BaseIV = Builder.createDerivedIV(
588 Kind, FPBinOp, StartV, CanonicalIV, Step, "offset.idx");
589
590 // Truncate base induction if needed.
591 VPTypeAnalysis TypeInfo(Plan);
592 Type *ResultTy = TypeInfo.inferScalarType(BaseIV);
593 if (TruncI) {
594 Type *TruncTy = TruncI->getType();
595 assert(ResultTy->getScalarSizeInBits() > TruncTy->getScalarSizeInBits() &&
596 "Not truncating.");
597 assert(ResultTy->isIntegerTy() && "Truncation requires an integer type");
598 BaseIV = Builder.createScalarCast(Instruction::Trunc, BaseIV, TruncTy, DL);
599 ResultTy = TruncTy;
600 }
601
602 // Truncate step if needed.
603 Type *StepTy = TypeInfo.inferScalarType(Step);
604 if (ResultTy != StepTy) {
605 assert(StepTy->getScalarSizeInBits() > ResultTy->getScalarSizeInBits() &&
606 "Not truncating.");
607 assert(StepTy->isIntegerTy() && "Truncation requires an integer type");
608 auto *VecPreheader =
610 VPBuilder::InsertPointGuard Guard(Builder);
611 Builder.setInsertPoint(VecPreheader);
612 Step = Builder.createScalarCast(Instruction::Trunc, Step, ResultTy, DL);
613 }
614 return Builder.createScalarIVSteps(InductionOpcode, FPBinOp, BaseIV, Step,
615 &Plan.getVF(), DL);
616}
617
620 for (unsigned I = 0; I != Users.size(); ++I) {
622 if (isa<VPHeaderPHIRecipe>(Cur))
623 continue;
624 for (VPValue *V : Cur->definedValues())
625 Users.insert_range(V->users());
626 }
627 return Users.takeVector();
628}
629
630/// Legalize VPWidenPointerInductionRecipe, by replacing it with a PtrAdd
631/// (IndStart, ScalarIVSteps (0, Step)) if only its scalar values are used, as
632/// VPWidenPointerInductionRecipe will generate vectors only. If some users
633/// require vectors while other require scalars, the scalar uses need to extract
634/// the scalars from the generated vectors (Note that this is different to how
635/// int/fp inductions are handled). Legalize extract-from-ends using uniform
636/// VPReplicateRecipe of wide inductions to use regular VPReplicateRecipe, so
637/// the correct end value is available. Also optimize
638/// VPWidenIntOrFpInductionRecipe, if any of its users needs scalar values, by
639/// providing them scalar steps built on the canonical scalar IV and update the
640/// original IV's users. This is an optional optimization to reduce the needs of
641/// vector extracts.
644 bool HasOnlyVectorVFs = !Plan.hasScalarVFOnly();
645 VPBuilder Builder(HeaderVPBB, HeaderVPBB->getFirstNonPhi());
646 for (VPRecipeBase &Phi : HeaderVPBB->phis()) {
647 auto *PhiR = dyn_cast<VPWidenInductionRecipe>(&Phi);
648 if (!PhiR)
649 continue;
650
651 // Try to narrow wide and replicating recipes to uniform recipes, based on
652 // VPlan analysis.
653 // TODO: Apply to all recipes in the future, to replace legacy uniformity
654 // analysis.
655 auto Users = collectUsersRecursively(PhiR);
656 for (VPUser *U : reverse(Users)) {
657 auto *Def = dyn_cast<VPSingleDefRecipe>(U);
658 auto *RepR = dyn_cast<VPReplicateRecipe>(U);
659 // Skip recipes that shouldn't be narrowed.
660 if (!Def || !isa<VPReplicateRecipe, VPWidenRecipe>(Def) ||
661 Def->getNumUsers() == 0 || !Def->getUnderlyingValue() ||
662 (RepR && (RepR->isSingleScalar() || RepR->isPredicated())))
663 continue;
664
665 // Skip recipes that may have other lanes than their first used.
667 continue;
668
669 auto *Clone = new VPReplicateRecipe(Def->getUnderlyingInstr(),
670 Def->operands(), /*IsUniform*/ true);
671 Clone->insertAfter(Def);
672 Def->replaceAllUsesWith(Clone);
673 }
674
675 // Replace wide pointer inductions which have only their scalars used by
676 // PtrAdd(IndStart, ScalarIVSteps (0, Step)).
677 if (auto *PtrIV = dyn_cast<VPWidenPointerInductionRecipe>(&Phi)) {
678 if (!PtrIV->onlyScalarsGenerated(Plan.hasScalableVF()))
679 continue;
680
681 const InductionDescriptor &ID = PtrIV->getInductionDescriptor();
682 VPValue *StartV =
683 Plan.getOrAddLiveIn(ConstantInt::get(ID.getStep()->getType(), 0));
684 VPValue *StepV = PtrIV->getOperand(1);
686 Plan, InductionDescriptor::IK_IntInduction, Instruction::Add, nullptr,
687 nullptr, StartV, StepV, PtrIV->getDebugLoc(), Builder);
688
689 VPValue *PtrAdd = Builder.createPtrAdd(PtrIV->getStartValue(), Steps,
690 PtrIV->getDebugLoc(), "next.gep");
691
692 PtrIV->replaceAllUsesWith(PtrAdd);
693 continue;
694 }
695
696 // Replace widened induction with scalar steps for users that only use
697 // scalars.
698 auto *WideIV = cast<VPWidenIntOrFpInductionRecipe>(&Phi);
699 if (HasOnlyVectorVFs && none_of(WideIV->users(), [WideIV](VPUser *U) {
700 return U->usesScalars(WideIV);
701 }))
702 continue;
703
704 const InductionDescriptor &ID = WideIV->getInductionDescriptor();
706 Plan, ID.getKind(), ID.getInductionOpcode(),
707 dyn_cast_or_null<FPMathOperator>(ID.getInductionBinOp()),
708 WideIV->getTruncInst(), WideIV->getStartValue(), WideIV->getStepValue(),
709 WideIV->getDebugLoc(), Builder);
710
711 // Update scalar users of IV to use Step instead.
712 if (!HasOnlyVectorVFs)
713 WideIV->replaceAllUsesWith(Steps);
714 else
715 WideIV->replaceUsesWithIf(Steps, [WideIV](VPUser &U, unsigned) {
716 return U.usesScalars(WideIV);
717 });
718 }
719}
720
721/// Check if \p VPV is an untruncated wide induction, either before or after the
722/// increment. If so return the header IV (before the increment), otherwise
723/// return null.
725 ScalarEvolution &SE) {
726 auto *WideIV = dyn_cast<VPWidenInductionRecipe>(VPV);
727 if (WideIV) {
728 // VPV itself is a wide induction, separately compute the end value for exit
729 // users if it is not a truncated IV.
730 auto *IntOrFpIV = dyn_cast<VPWidenIntOrFpInductionRecipe>(WideIV);
731 return (IntOrFpIV && IntOrFpIV->getTruncInst()) ? nullptr : WideIV;
732 }
733
734 // Check if VPV is an optimizable induction increment.
735 VPRecipeBase *Def = VPV->getDefiningRecipe();
736 if (!Def || Def->getNumOperands() != 2)
737 return nullptr;
738 WideIV = dyn_cast<VPWidenInductionRecipe>(Def->getOperand(0));
739 if (!WideIV)
740 WideIV = dyn_cast<VPWidenInductionRecipe>(Def->getOperand(1));
741 if (!WideIV)
742 return nullptr;
743
744 auto IsWideIVInc = [&]() {
745 auto &ID = WideIV->getInductionDescriptor();
746
747 // Check if VPV increments the induction by the induction step.
748 VPValue *IVStep = WideIV->getStepValue();
749 switch (ID.getInductionOpcode()) {
750 case Instruction::Add:
751 return match(VPV, m_c_Add(m_Specific(WideIV), m_Specific(IVStep)));
752 case Instruction::FAdd:
754 m_Specific(IVStep)));
755 case Instruction::FSub:
756 return match(VPV, m_Binary<Instruction::FSub>(m_Specific(WideIV),
757 m_Specific(IVStep)));
758 case Instruction::Sub: {
759 // IVStep will be the negated step of the subtraction. Check if Step == -1
760 // * IVStep.
761 VPValue *Step;
762 if (!match(VPV, m_Sub(m_VPValue(), m_VPValue(Step))))
763 return false;
764 const SCEV *IVStepSCEV = vputils::getSCEVExprForVPValue(IVStep, SE);
765 const SCEV *StepSCEV = vputils::getSCEVExprForVPValue(Step, SE);
766 return !isa<SCEVCouldNotCompute>(IVStepSCEV) &&
767 !isa<SCEVCouldNotCompute>(StepSCEV) &&
768 IVStepSCEV == SE.getNegativeSCEV(StepSCEV);
769 }
770 default:
771 return ID.getKind() == InductionDescriptor::IK_PtrInduction &&
772 match(VPV, m_GetElementPtr(m_Specific(WideIV),
773 m_Specific(WideIV->getStepValue())));
774 }
775 llvm_unreachable("should have been covered by switch above");
776 };
777 return IsWideIVInc() ? WideIV : nullptr;
778}
779
780/// Attempts to optimize the induction variable exit values for users in the
781/// early exit block.
783 VPTypeAnalysis &TypeInfo,
784 VPBlockBase *PredVPBB,
785 VPValue *Op,
786 ScalarEvolution &SE) {
787 VPValue *Incoming, *Mask;
790 m_VPValue(Mask)),
792 return nullptr;
793
794 auto *WideIV = getOptimizableIVOf(Incoming, SE);
795 if (!WideIV)
796 return nullptr;
797
798 auto *WideIntOrFp = dyn_cast<VPWidenIntOrFpInductionRecipe>(WideIV);
799 if (WideIntOrFp && WideIntOrFp->getTruncInst())
800 return nullptr;
801
802 // Calculate the final index.
803 VPValue *EndValue = Plan.getCanonicalIV();
804 auto CanonicalIVType = Plan.getCanonicalIV()->getScalarType();
805 VPBuilder B(cast<VPBasicBlock>(PredVPBB));
806
807 DebugLoc DL = cast<VPInstruction>(Op)->getDebugLoc();
808 VPValue *FirstActiveLane =
809 B.createNaryOp(VPInstruction::FirstActiveLane, Mask, DL);
810 Type *FirstActiveLaneType = TypeInfo.inferScalarType(FirstActiveLane);
811 FirstActiveLane = B.createScalarZExtOrTrunc(FirstActiveLane, CanonicalIVType,
812 FirstActiveLaneType, DL);
813 EndValue = B.createNaryOp(Instruction::Add, {EndValue, FirstActiveLane}, DL);
814
815 // `getOptimizableIVOf()` always returns the pre-incremented IV, so if it
816 // changed it means the exit is using the incremented value, so we need to
817 // add the step.
818 if (Incoming != WideIV) {
819 VPValue *One = Plan.getOrAddLiveIn(ConstantInt::get(CanonicalIVType, 1));
820 EndValue = B.createNaryOp(Instruction::Add, {EndValue, One}, DL);
821 }
822
823 if (!WideIntOrFp || !WideIntOrFp->isCanonical()) {
824 const InductionDescriptor &ID = WideIV->getInductionDescriptor();
825 VPValue *Start = WideIV->getStartValue();
826 VPValue *Step = WideIV->getStepValue();
827 EndValue = B.createDerivedIV(
828 ID.getKind(), dyn_cast_or_null<FPMathOperator>(ID.getInductionBinOp()),
829 Start, EndValue, Step);
830 }
831
832 return EndValue;
833}
834
835/// Attempts to optimize the induction variable exit values for users in the
836/// exit block coming from the latch in the original scalar loop.
838 VPlan &Plan, VPTypeAnalysis &TypeInfo, VPBlockBase *PredVPBB, VPValue *Op,
842 return nullptr;
843
844 auto *WideIV = getOptimizableIVOf(Incoming, SE);
845 if (!WideIV)
846 return nullptr;
847
848 VPValue *EndValue = EndValues.lookup(WideIV);
849 assert(EndValue && "end value must have been pre-computed");
850
851 // `getOptimizableIVOf()` always returns the pre-incremented IV, so if it
852 // changed it means the exit is using the incremented value, so we don't
853 // need to subtract the step.
854 if (Incoming != WideIV)
855 return EndValue;
856
857 // Otherwise, subtract the step from the EndValue.
858 VPBuilder B(cast<VPBasicBlock>(PredVPBB)->getTerminator());
859 VPValue *Step = WideIV->getStepValue();
860 Type *ScalarTy = TypeInfo.inferScalarType(WideIV);
861 if (ScalarTy->isIntegerTy())
862 return B.createNaryOp(Instruction::Sub, {EndValue, Step}, {}, "ind.escape");
863 if (ScalarTy->isPointerTy()) {
864 Type *StepTy = TypeInfo.inferScalarType(Step);
865 auto *Zero = Plan.getOrAddLiveIn(ConstantInt::get(StepTy, 0));
866 return B.createPtrAdd(EndValue,
867 B.createNaryOp(Instruction::Sub, {Zero, Step}),
868 DebugLoc::getUnknown(), "ind.escape");
869 }
870 if (ScalarTy->isFloatingPointTy()) {
871 const auto &ID = WideIV->getInductionDescriptor();
872 return B.createNaryOp(
873 ID.getInductionBinOp()->getOpcode() == Instruction::FAdd
874 ? Instruction::FSub
875 : Instruction::FAdd,
876 {EndValue, Step}, {ID.getInductionBinOp()->getFastMathFlags()});
877 }
878 llvm_unreachable("all possible induction types must be handled");
879 return nullptr;
880}
881
883 VPlan &Plan, DenseMap<VPValue *, VPValue *> &EndValues,
884 ScalarEvolution &SE) {
885 VPBlockBase *MiddleVPBB = Plan.getMiddleBlock();
886 VPTypeAnalysis TypeInfo(Plan);
887 for (VPIRBasicBlock *ExitVPBB : Plan.getExitBlocks()) {
888 for (VPRecipeBase &R : ExitVPBB->phis()) {
889 auto *ExitIRI = cast<VPIRPhi>(&R);
890
891 for (auto [Idx, PredVPBB] : enumerate(ExitVPBB->getPredecessors())) {
892 VPValue *Escape = nullptr;
893 if (PredVPBB == MiddleVPBB)
894 Escape = optimizeLatchExitInductionUser(Plan, TypeInfo, PredVPBB,
895 ExitIRI->getOperand(Idx),
896 EndValues, SE);
897 else
898 Escape = optimizeEarlyExitInductionUser(Plan, TypeInfo, PredVPBB,
899 ExitIRI->getOperand(Idx), SE);
900 if (Escape)
901 ExitIRI->setOperand(Idx, Escape);
902 }
903 }
904 }
905}
906
907/// Remove redundant EpxandSCEVRecipes in \p Plan's entry block by replacing
908/// them with already existing recipes expanding the same SCEV expression.
911
912 for (VPRecipeBase &R :
914 auto *ExpR = dyn_cast<VPExpandSCEVRecipe>(&R);
915 if (!ExpR)
916 continue;
917
918 const auto &[V, Inserted] = SCEV2VPV.try_emplace(ExpR->getSCEV(), ExpR);
919 if (Inserted)
920 continue;
921 ExpR->replaceAllUsesWith(V->second);
922 ExpR->eraseFromParent();
923 }
924}
925
927 SmallVector<VPValue *> WorkList;
929 WorkList.push_back(V);
930
931 while (!WorkList.empty()) {
932 VPValue *Cur = WorkList.pop_back_val();
933 if (!Seen.insert(Cur).second)
934 continue;
936 if (!R)
937 continue;
938 if (!isDeadRecipe(*R))
939 continue;
940 WorkList.append(R->op_begin(), R->op_end());
941 R->eraseFromParent();
942 }
943}
944
945/// Try to fold \p R using InstSimplifyFolder. Will succeed and return a
946/// non-nullptr Value for a handled \p Opcode if corresponding \p Operands are
947/// foldable live-ins.
948static Value *tryToFoldLiveIns(const VPRecipeBase &R, unsigned Opcode,
950 const DataLayout &DL, VPTypeAnalysis &TypeInfo) {
952 for (VPValue *Op : Operands) {
953 if (!Op->isLiveIn() || !Op->getLiveInIRValue())
954 return nullptr;
955 Ops.push_back(Op->getLiveInIRValue());
956 }
957
958 InstSimplifyFolder Folder(DL);
959 if (Instruction::isBinaryOp(Opcode))
960 return Folder.FoldBinOp(static_cast<Instruction::BinaryOps>(Opcode), Ops[0],
961 Ops[1]);
962 if (Instruction::isCast(Opcode))
963 return Folder.FoldCast(static_cast<Instruction::CastOps>(Opcode), Ops[0],
964 TypeInfo.inferScalarType(R.getVPSingleValue()));
965 switch (Opcode) {
967 return Folder.FoldSelect(Ops[0], Ops[1],
970 return Folder.FoldBinOp(Instruction::BinaryOps::Xor, Ops[0],
972 case Instruction::Select:
973 return Folder.FoldSelect(Ops[0], Ops[1], Ops[2]);
974 case Instruction::ICmp:
975 case Instruction::FCmp:
976 return Folder.FoldCmp(cast<VPRecipeWithIRFlags>(R).getPredicate(), Ops[0],
977 Ops[1]);
978 case Instruction::GetElementPtr: {
979 auto &RFlags = cast<VPRecipeWithIRFlags>(R);
980 auto *GEP = cast<GetElementPtrInst>(RFlags.getUnderlyingInstr());
981 return Folder.FoldGEP(GEP->getSourceElementType(), Ops[0], drop_begin(Ops),
982 RFlags.getGEPNoWrapFlags());
983 }
986 return Folder.FoldGEP(IntegerType::getInt8Ty(TypeInfo.getContext()), Ops[0],
987 Ops[1],
988 cast<VPRecipeWithIRFlags>(R).getGEPNoWrapFlags());
989 // An extract of a live-in is an extract of a broadcast, so return the
990 // broadcasted element.
991 case Instruction::ExtractElement:
992 assert(!Ops[0]->getType()->isVectorTy() && "Live-ins should be scalar");
993 return Ops[0];
994 }
995 return nullptr;
996}
997
998/// Try to simplify recipe \p R.
999static void simplifyRecipe(VPRecipeBase &R, VPTypeAnalysis &TypeInfo) {
1000 VPlan *Plan = R.getParent()->getPlan();
1001
1002 auto *Def = dyn_cast<VPSingleDefRecipe>(&R);
1003 if (!Def)
1004 return;
1005
1006 // Simplification of live-in IR values for SingleDef recipes using
1007 // InstSimplifyFolder.
1011 const DataLayout &DL =
1013 Value *V = tryToFoldLiveIns(*I, I->getOpcode(), I->operands(), DL,
1014 TypeInfo);
1015 if (V)
1016 I->replaceAllUsesWith(Plan->getOrAddLiveIn(V));
1017 return V;
1018 })
1019 .Default([](auto *) { return false; }))
1020 return;
1021
1022 // Fold PredPHI LiveIn -> LiveIn.
1023 if (auto *PredPHI = dyn_cast<VPPredInstPHIRecipe>(&R)) {
1024 VPValue *Op = PredPHI->getOperand(0);
1025 if (Op->isLiveIn())
1026 PredPHI->replaceAllUsesWith(Op);
1027 }
1028
1029 VPValue *A;
1030 if (match(Def, m_Trunc(m_ZExtOrSExt(m_VPValue(A))))) {
1031 Type *TruncTy = TypeInfo.inferScalarType(Def);
1032 Type *ATy = TypeInfo.inferScalarType(A);
1033 if (TruncTy == ATy) {
1034 Def->replaceAllUsesWith(A);
1035 } else {
1036 // Don't replace a scalarizing recipe with a widened cast.
1037 if (isa<VPReplicateRecipe>(Def))
1038 return;
1039 if (ATy->getScalarSizeInBits() < TruncTy->getScalarSizeInBits()) {
1040
1041 unsigned ExtOpcode = match(R.getOperand(0), m_SExt(m_VPValue()))
1042 ? Instruction::SExt
1043 : Instruction::ZExt;
1044 auto *VPC =
1045 new VPWidenCastRecipe(Instruction::CastOps(ExtOpcode), A, TruncTy);
1046 if (auto *UnderlyingExt = R.getOperand(0)->getUnderlyingValue()) {
1047 // UnderlyingExt has distinct return type, used to retain legacy cost.
1048 VPC->setUnderlyingValue(UnderlyingExt);
1049 }
1050 VPC->insertBefore(&R);
1051 Def->replaceAllUsesWith(VPC);
1052 } else if (ATy->getScalarSizeInBits() > TruncTy->getScalarSizeInBits()) {
1053 auto *VPC = new VPWidenCastRecipe(Instruction::Trunc, A, TruncTy);
1054 VPC->insertBefore(&R);
1055 Def->replaceAllUsesWith(VPC);
1056 }
1057 }
1058#ifndef NDEBUG
1059 // Verify that the cached type info is for both A and its users is still
1060 // accurate by comparing it to freshly computed types.
1061 VPTypeAnalysis TypeInfo2(*Plan);
1062 assert(TypeInfo.inferScalarType(A) == TypeInfo2.inferScalarType(A));
1063 for (VPUser *U : A->users()) {
1064 auto *R = cast<VPRecipeBase>(U);
1065 for (VPValue *VPV : R->definedValues())
1066 assert(TypeInfo.inferScalarType(VPV) == TypeInfo2.inferScalarType(VPV));
1067 }
1068#endif
1069 }
1070
1071 // Simplify (X && Y) || (X && !Y) -> X.
1072 // TODO: Split up into simpler, modular combines: (X && Y) || (X && Z) into X
1073 // && (Y || Z) and (X || !X) into true. This requires queuing newly created
1074 // recipes to be visited during simplification.
1075 VPValue *X, *Y, *Z;
1076 if (match(Def,
1079 Def->replaceAllUsesWith(X);
1080 Def->eraseFromParent();
1081 return;
1082 }
1083
1084 // x | 1 -> 1
1085 if (match(Def, m_c_BinaryOr(m_VPValue(X), m_AllOnes())))
1086 return Def->replaceAllUsesWith(Def->getOperand(Def->getOperand(0) == X));
1087
1088 // x | 0 -> x
1089 if (match(Def, m_c_BinaryOr(m_VPValue(X), m_ZeroInt())))
1090 return Def->replaceAllUsesWith(X);
1091
1092 // x & 0 -> 0
1093 if (match(Def, m_c_BinaryAnd(m_VPValue(X), m_ZeroInt())))
1094 return Def->replaceAllUsesWith(Def->getOperand(Def->getOperand(0) == X));
1095
1096 // x && false -> false
1097 if (match(Def, m_LogicalAnd(m_VPValue(X), m_False())))
1098 return Def->replaceAllUsesWith(Def->getOperand(1));
1099
1100 // (x && y) || (x && z) -> x && (y || z)
1101 VPBuilder Builder(Def);
1104 // Simplify only if one of the operands has one use to avoid creating an
1105 // extra recipe.
1106 (!Def->getOperand(0)->hasMoreThanOneUniqueUser() ||
1107 !Def->getOperand(1)->hasMoreThanOneUniqueUser()))
1108 return Def->replaceAllUsesWith(
1109 Builder.createLogicalAnd(X, Builder.createOr(Y, Z)));
1110
1111 // x && !x -> 0
1113 return Def->replaceAllUsesWith(Plan->getOrAddLiveIn(
1115
1116 if (match(Def, m_Select(m_VPValue(), m_VPValue(X), m_Deferred(X))))
1117 return Def->replaceAllUsesWith(X);
1118
1119 // select !c, x, y -> select c, y, x
1120 VPValue *C;
1121 if (match(Def, m_Select(m_Not(m_VPValue(C)), m_VPValue(X), m_VPValue(Y)))) {
1122 Def->setOperand(0, C);
1123 Def->setOperand(1, Y);
1124 Def->setOperand(2, X);
1125 return;
1126 }
1127
1128 // Reassociate (x && y) && z -> x && (y && z) if x has multiple users. With
1129 // tail folding it is likely that x is a header mask and can be simplified
1130 // further.
1132 m_VPValue(Z))) &&
1133 X->hasMoreThanOneUniqueUser())
1134 return Def->replaceAllUsesWith(
1135 Builder.createLogicalAnd(X, Builder.createLogicalAnd(Y, Z)));
1136
1137 if (match(Def, m_c_Mul(m_VPValue(A), m_One())))
1138 return Def->replaceAllUsesWith(A);
1139
1140 if (match(Def, m_c_Mul(m_VPValue(A), m_ZeroInt())))
1141 return Def->replaceAllUsesWith(R.getOperand(0) == A ? R.getOperand(1)
1142 : R.getOperand(0));
1143
1144 if (match(Def, m_Not(m_VPValue(A)))) {
1145 if (match(A, m_Not(m_VPValue(A))))
1146 return Def->replaceAllUsesWith(A);
1147
1148 // Try to fold Not into compares by adjusting the predicate in-place.
1149 CmpPredicate Pred;
1150 if (match(A, m_Cmp(Pred, m_VPValue(), m_VPValue()))) {
1151 auto *Cmp = cast<VPRecipeWithIRFlags>(A);
1152 if (all_of(Cmp->users(),
1154 m_Not(m_Specific(Cmp)),
1155 m_Select(m_Specific(Cmp), m_VPValue(), m_VPValue()))))) {
1156 Cmp->setPredicate(CmpInst::getInversePredicate(Pred));
1157 for (VPUser *U : to_vector(Cmp->users())) {
1158 auto *R = cast<VPSingleDefRecipe>(U);
1159 if (match(R, m_Select(m_Specific(Cmp), m_VPValue(X), m_VPValue(Y)))) {
1160 // select (cmp pred), x, y -> select (cmp inv_pred), y, x
1161 R->setOperand(1, Y);
1162 R->setOperand(2, X);
1163 } else {
1164 // not (cmp pred) -> cmp inv_pred
1165 assert(match(R, m_Not(m_Specific(Cmp))) && "Unexpected user");
1166 R->replaceAllUsesWith(Cmp);
1167 }
1168 }
1169 // If Cmp doesn't have a debug location, use the one from the negation,
1170 // to preserve the location.
1171 if (!Cmp->getDebugLoc() && R.getDebugLoc())
1172 Cmp->setDebugLoc(R.getDebugLoc());
1173 }
1174 }
1175 }
1176
1177 // Remove redundant DerviedIVs, that is 0 + A * 1 -> A and 0 + 0 * x -> 0.
1178 if ((match(Def, m_DerivedIV(m_ZeroInt(), m_VPValue(A), m_One())) ||
1179 match(Def, m_DerivedIV(m_ZeroInt(), m_ZeroInt(), m_VPValue()))) &&
1180 TypeInfo.inferScalarType(Def->getOperand(1)) ==
1181 TypeInfo.inferScalarType(Def))
1182 return Def->replaceAllUsesWith(Def->getOperand(1));
1183
1185 m_One()))) {
1186 Type *WideStepTy = TypeInfo.inferScalarType(Def);
1187 if (TypeInfo.inferScalarType(X) != WideStepTy)
1188 X = Builder.createWidenCast(Instruction::Trunc, X, WideStepTy);
1189 Def->replaceAllUsesWith(X);
1190 return;
1191 }
1192
1193 // For i1 vp.merges produced by AnyOf reductions:
1194 // vp.merge true, (or x, y), x, evl -> vp.merge y, true, x, evl
1196 m_VPValue(X), m_VPValue())) &&
1198 TypeInfo.inferScalarType(R.getVPSingleValue())->isIntegerTy(1)) {
1199 Def->setOperand(1, Def->getOperand(0));
1200 Def->setOperand(0, Y);
1201 return;
1202 }
1203
1204 if (auto *Phi = dyn_cast<VPFirstOrderRecurrencePHIRecipe>(Def)) {
1205 if (Phi->getOperand(0) == Phi->getOperand(1))
1206 Def->replaceAllUsesWith(Phi->getOperand(0));
1207 return;
1208 }
1209
1210 // Look through ExtractLastElement (BuildVector ....).
1212 auto *BuildVector = cast<VPInstruction>(R.getOperand(0));
1213 Def->replaceAllUsesWith(
1214 BuildVector->getOperand(BuildVector->getNumOperands() - 1));
1215 return;
1216 }
1217
1218 // Look through ExtractPenultimateElement (BuildVector ....).
1220 m_BuildVector()))) {
1221 auto *BuildVector = cast<VPInstruction>(R.getOperand(0));
1222 Def->replaceAllUsesWith(
1223 BuildVector->getOperand(BuildVector->getNumOperands() - 2));
1224 return;
1225 }
1226
1227 if (auto *Phi = dyn_cast<VPPhi>(Def)) {
1228 if (Phi->getNumOperands() == 1)
1229 Phi->replaceAllUsesWith(Phi->getOperand(0));
1230 return;
1231 }
1232
1233 // Some simplifications can only be applied after unrolling. Perform them
1234 // below.
1235 if (!Plan->isUnrolled())
1236 return;
1237
1238 // VPVectorPointer for part 0 can be replaced by their start pointer.
1239 if (auto *VecPtr = dyn_cast<VPVectorPointerRecipe>(&R)) {
1240 if (VecPtr->isFirstPart()) {
1241 VecPtr->replaceAllUsesWith(VecPtr->getOperand(0));
1242 return;
1243 }
1244 }
1245
1246 // VPScalarIVSteps for part 0 can be replaced by their start value, if only
1247 // the first lane is demanded.
1248 if (auto *Steps = dyn_cast<VPScalarIVStepsRecipe>(Def)) {
1249 if (Steps->isPart0() && vputils::onlyFirstLaneUsed(Steps)) {
1250 Steps->replaceAllUsesWith(Steps->getOperand(0));
1251 return;
1252 }
1253 }
1254 // Simplify redundant ReductionStartVector recipes after unrolling.
1255 VPValue *StartV;
1257 m_VPValue(StartV), m_VPValue(), m_VPValue()))) {
1258 Def->replaceUsesWithIf(StartV, [](const VPUser &U, unsigned Idx) {
1259 auto *PhiR = dyn_cast<VPReductionPHIRecipe>(&U);
1260 return PhiR && PhiR->isInLoop();
1261 });
1262 return;
1263 }
1264
1266 Def->replaceAllUsesWith(A);
1267 return;
1268 }
1269
1270 if (match(Def,
1274 cast<VPReplicateRecipe>(A)->isSingleScalar())) &&
1275 all_of(A->users(),
1276 [Def, A](VPUser *U) { return U->usesScalars(A) || Def == U; })) {
1277 return Def->replaceAllUsesWith(A);
1278 }
1279}
1280
1283 Plan.getEntry());
1284 VPTypeAnalysis TypeInfo(Plan);
1286 for (VPRecipeBase &R : make_early_inc_range(*VPBB)) {
1287 simplifyRecipe(R, TypeInfo);
1288 }
1289 }
1290}
1291
1293 if (Plan.hasScalarVFOnly())
1294 return;
1295
1296 // Try to narrow wide and replicating recipes to single scalar recipes,
1297 // based on VPlan analysis. Only process blocks in the loop region for now,
1298 // without traversing into nested regions, as recipes in replicate regions
1299 // cannot be converted yet.
1302 for (VPRecipeBase &R : make_early_inc_range(reverse(*VPBB))) {
1304 continue;
1305 auto *RepR = dyn_cast<VPReplicateRecipe>(&R);
1306 if (RepR && (RepR->isSingleScalar() || RepR->isPredicated()))
1307 continue;
1308
1309 auto *RepOrWidenR = cast<VPSingleDefRecipe>(&R);
1310 if (RepR && isa<StoreInst>(RepR->getUnderlyingInstr()) &&
1311 vputils::isSingleScalar(RepR->getOperand(1))) {
1312 auto *Clone = new VPReplicateRecipe(
1313 RepOrWidenR->getUnderlyingInstr(), RepOrWidenR->operands(),
1314 true /*IsSingleScalar*/, nullptr /*Mask*/, *RepR /*Metadata*/);
1315 Clone->insertBefore(RepOrWidenR);
1317 {Clone->getOperand(0)});
1318 Ext->insertBefore(Clone);
1319 Clone->setOperand(0, Ext);
1320 RepR->eraseFromParent();
1321 continue;
1322 }
1323
1324 // Skip recipes that aren't single scalars or don't have only their
1325 // scalar results used. In the latter case, we would introduce extra
1326 // broadcasts.
1327 if (!vputils::isSingleScalar(RepOrWidenR) ||
1328 !all_of(RepOrWidenR->users(), [RepOrWidenR](const VPUser *U) {
1329 return U->usesScalars(RepOrWidenR) ||
1330 match(cast<VPRecipeBase>(U),
1331 m_ExtractLastElement(m_VPValue()));
1332 }))
1333 continue;
1334
1335 auto *Clone = new VPReplicateRecipe(RepOrWidenR->getUnderlyingInstr(),
1336 RepOrWidenR->operands(),
1337 true /*IsSingleScalar*/);
1338 Clone->insertBefore(RepOrWidenR);
1339 RepOrWidenR->replaceAllUsesWith(Clone);
1340 }
1341 }
1342}
1343
1344/// Try to see if all of \p Blend's masks share a common value logically and'ed
1345/// and remove it from the masks.
1347 if (Blend->isNormalized())
1348 return;
1349 VPValue *CommonEdgeMask;
1350 if (!match(Blend->getMask(0),
1351 m_LogicalAnd(m_VPValue(CommonEdgeMask), m_VPValue())))
1352 return;
1353 for (unsigned I = 0; I < Blend->getNumIncomingValues(); I++)
1354 if (!match(Blend->getMask(I),
1355 m_LogicalAnd(m_Specific(CommonEdgeMask), m_VPValue())))
1356 return;
1357 for (unsigned I = 0; I < Blend->getNumIncomingValues(); I++)
1358 Blend->setMask(I, Blend->getMask(I)->getDefiningRecipe()->getOperand(1));
1359}
1360
1361/// Normalize and simplify VPBlendRecipes. Should be run after simplifyRecipes
1362/// to make sure the masks are simplified.
1363static void simplifyBlends(VPlan &Plan) {
1366 for (VPRecipeBase &R : make_early_inc_range(*VPBB)) {
1367 auto *Blend = dyn_cast<VPBlendRecipe>(&R);
1368 if (!Blend)
1369 continue;
1370
1371 removeCommonBlendMask(Blend);
1372
1373 // Try to remove redundant blend recipes.
1374 SmallPtrSet<VPValue *, 4> UniqueValues;
1375 if (Blend->isNormalized() || !match(Blend->getMask(0), m_False()))
1376 UniqueValues.insert(Blend->getIncomingValue(0));
1377 for (unsigned I = 1; I != Blend->getNumIncomingValues(); ++I)
1378 if (!match(Blend->getMask(I), m_False()))
1379 UniqueValues.insert(Blend->getIncomingValue(I));
1380
1381 if (UniqueValues.size() == 1) {
1382 Blend->replaceAllUsesWith(*UniqueValues.begin());
1383 Blend->eraseFromParent();
1384 continue;
1385 }
1386
1387 if (Blend->isNormalized())
1388 continue;
1389
1390 // Normalize the blend so its first incoming value is used as the initial
1391 // value with the others blended into it.
1392
1393 unsigned StartIndex = 0;
1394 for (unsigned I = 0; I != Blend->getNumIncomingValues(); ++I) {
1395 // If a value's mask is used only by the blend then is can be deadcoded.
1396 // TODO: Find the most expensive mask that can be deadcoded, or a mask
1397 // that's used by multiple blends where it can be removed from them all.
1398 VPValue *Mask = Blend->getMask(I);
1399 if (Mask->getNumUsers() == 1 && !match(Mask, m_False())) {
1400 StartIndex = I;
1401 break;
1402 }
1403 }
1404
1405 SmallVector<VPValue *, 4> OperandsWithMask;
1406 OperandsWithMask.push_back(Blend->getIncomingValue(StartIndex));
1407
1408 for (unsigned I = 0; I != Blend->getNumIncomingValues(); ++I) {
1409 if (I == StartIndex)
1410 continue;
1411 OperandsWithMask.push_back(Blend->getIncomingValue(I));
1412 OperandsWithMask.push_back(Blend->getMask(I));
1413 }
1414
1415 auto *NewBlend =
1416 new VPBlendRecipe(cast_or_null<PHINode>(Blend->getUnderlyingValue()),
1417 OperandsWithMask, Blend->getDebugLoc());
1418 NewBlend->insertBefore(&R);
1419
1420 VPValue *DeadMask = Blend->getMask(StartIndex);
1421 Blend->replaceAllUsesWith(NewBlend);
1422 Blend->eraseFromParent();
1424
1425 /// Simplify BLEND %a, %b, Not(%mask) -> BLEND %b, %a, %mask.
1426 VPValue *NewMask;
1427 if (NewBlend->getNumOperands() == 3 &&
1428 match(NewBlend->getMask(1), m_Not(m_VPValue(NewMask)))) {
1429 VPValue *Inc0 = NewBlend->getOperand(0);
1430 VPValue *Inc1 = NewBlend->getOperand(1);
1431 VPValue *OldMask = NewBlend->getOperand(2);
1432 NewBlend->setOperand(0, Inc1);
1433 NewBlend->setOperand(1, Inc0);
1434 NewBlend->setOperand(2, NewMask);
1435 if (OldMask->getNumUsers() == 0)
1436 cast<VPInstruction>(OldMask)->eraseFromParent();
1437 }
1438 }
1439 }
1440}
1441
1442/// Optimize the width of vector induction variables in \p Plan based on a known
1443/// constant Trip Count, \p BestVF and \p BestUF.
1445 ElementCount BestVF,
1446 unsigned BestUF) {
1447 // Only proceed if we have not completely removed the vector region.
1448 if (!Plan.getVectorLoopRegion())
1449 return false;
1450
1451 if (!Plan.getTripCount()->isLiveIn())
1452 return false;
1455 if (!TC || !BestVF.isFixed())
1456 return false;
1457
1458 // Calculate the minimum power-of-2 bit width that can fit the known TC, VF
1459 // and UF. Returns at least 8.
1460 auto ComputeBitWidth = [](APInt TC, uint64_t Align) {
1461 APInt AlignedTC =
1464 APInt MaxVal = AlignedTC - 1;
1465 return std::max<unsigned>(PowerOf2Ceil(MaxVal.getActiveBits()), 8);
1466 };
1467 unsigned NewBitWidth =
1468 ComputeBitWidth(TC->getValue(), BestVF.getKnownMinValue() * BestUF);
1469
1470 LLVMContext &Ctx = Plan.getContext();
1471 auto *NewIVTy = IntegerType::get(Ctx, NewBitWidth);
1472
1473 bool MadeChange = false;
1474
1475 VPBasicBlock *HeaderVPBB = Plan.getVectorLoopRegion()->getEntryBasicBlock();
1476 for (VPRecipeBase &Phi : HeaderVPBB->phis()) {
1477 auto *WideIV = dyn_cast<VPWidenIntOrFpInductionRecipe>(&Phi);
1478
1479 // Currently only handle canonical IVs as it is trivial to replace the start
1480 // and stop values, and we currently only perform the optimization when the
1481 // IV has a single use.
1482 if (!WideIV || !WideIV->isCanonical() ||
1483 WideIV->hasMoreThanOneUniqueUser() ||
1484 NewIVTy == WideIV->getScalarType())
1485 continue;
1486
1487 // Currently only handle cases where the single user is a header-mask
1488 // comparison with the backedge-taken-count.
1489 if (!match(*WideIV->user_begin(),
1490 m_ICmp(m_Specific(WideIV),
1493 continue;
1494
1495 // Update IV operands and comparison bound to use new narrower type.
1496 auto *NewStart = Plan.getOrAddLiveIn(ConstantInt::get(NewIVTy, 0));
1497 WideIV->setStartValue(NewStart);
1498 auto *NewStep = Plan.getOrAddLiveIn(ConstantInt::get(NewIVTy, 1));
1499 WideIV->setStepValue(NewStep);
1500
1501 auto *NewBTC = new VPWidenCastRecipe(
1502 Instruction::Trunc, Plan.getOrCreateBackedgeTakenCount(), NewIVTy);
1503 Plan.getVectorPreheader()->appendRecipe(NewBTC);
1504 auto *Cmp = cast<VPInstruction>(*WideIV->user_begin());
1505 Cmp->setOperand(1, NewBTC);
1506
1507 MadeChange = true;
1508 }
1509
1510 return MadeChange;
1511}
1512
1513/// Return true if \p Cond is known to be true for given \p BestVF and \p
1514/// BestUF.
1516 ElementCount BestVF, unsigned BestUF,
1517 ScalarEvolution &SE) {
1519 return any_of(Cond->getDefiningRecipe()->operands(), [&Plan, BestVF, BestUF,
1520 &SE](VPValue *C) {
1521 return isConditionTrueViaVFAndUF(C, Plan, BestVF, BestUF, SE);
1522 });
1523
1524 auto *CanIV = Plan.getCanonicalIV();
1526 m_Specific(CanIV->getBackedgeValue()),
1527 m_Specific(&Plan.getVectorTripCount()))))
1528 return false;
1529
1530 // The compare checks CanIV + VFxUF == vector trip count. The vector trip
1531 // count is not conveniently available as SCEV so far, so we compare directly
1532 // against the original trip count. This is stricter than necessary, as we
1533 // will only return true if the trip count == vector trip count.
1534 const SCEV *VectorTripCount =
1536 if (isa<SCEVCouldNotCompute>(VectorTripCount))
1537 VectorTripCount = vputils::getSCEVExprForVPValue(Plan.getTripCount(), SE);
1538 assert(!isa<SCEVCouldNotCompute>(VectorTripCount) &&
1539 "Trip count SCEV must be computable");
1540 ElementCount NumElements = BestVF.multiplyCoefficientBy(BestUF);
1541 const SCEV *C = SE.getElementCount(VectorTripCount->getType(), NumElements);
1542 return SE.isKnownPredicate(CmpInst::ICMP_EQ, VectorTripCount, C);
1543}
1544
1545/// Try to replace multiple active lane masks used for control flow with
1546/// a single, wide active lane mask instruction followed by multiple
1547/// extract subvector intrinsics. This applies to the active lane mask
1548/// instructions both in the loop and in the preheader.
1549/// Incoming values of all ActiveLaneMaskPHIs are updated to use the
1550/// new extracts from the first active lane mask, which has it's last
1551/// operand (multiplier) set to UF.
1553 unsigned UF) {
1554 if (!EnableWideActiveLaneMask || !VF.isVector() || UF == 1)
1555 return false;
1556
1557 VPRegionBlock *VectorRegion = Plan.getVectorLoopRegion();
1558 VPBasicBlock *ExitingVPBB = VectorRegion->getExitingBasicBlock();
1559 auto *Term = &ExitingVPBB->back();
1560
1561 using namespace llvm::VPlanPatternMatch;
1563 m_VPValue(), m_VPValue(), m_VPValue())))))
1564 return false;
1565
1566 auto *Header = cast<VPBasicBlock>(VectorRegion->getEntry());
1567 LLVMContext &Ctx = Plan.getContext();
1568
1569 auto ExtractFromALM = [&](VPInstruction *ALM,
1570 SmallVectorImpl<VPValue *> &Extracts) {
1571 DebugLoc DL = ALM->getDebugLoc();
1572 for (unsigned Part = 0; Part < UF; ++Part) {
1574 Ops.append({ALM, Plan.getOrAddLiveIn(
1575 ConstantInt::get(IntegerType::getInt64Ty(Ctx),
1576 VF.getKnownMinValue() * Part))});
1577 auto *Ext = new VPWidenIntrinsicRecipe(Intrinsic::vector_extract, Ops,
1579 Extracts[Part] = Ext;
1580 Ext->insertAfter(ALM);
1581 }
1582 };
1583
1584 // Create a list of each active lane mask phi, ordered by unroll part.
1586 for (VPRecipeBase &R : Header->phis()) {
1588 if (!Phi)
1589 continue;
1590 VPValue *Index = nullptr;
1591 match(Phi->getBackedgeValue(),
1593 assert(Index && "Expected index from ActiveLaneMask instruction");
1594
1595 auto *II = dyn_cast<VPInstruction>(Index);
1596 if (II && II->getOpcode() == VPInstruction::CanonicalIVIncrementForPart) {
1597 auto Part = cast<ConstantInt>(II->getOperand(1)->getLiveInIRValue());
1598 Phis[Part->getZExtValue()] = Phi;
1599 } else
1600 // Anything other than a CanonicalIVIncrementForPart is part 0
1601 Phis[0] = Phi;
1602 }
1603
1604 assert(all_of(Phis, [](VPActiveLaneMaskPHIRecipe *Phi) { return Phi; }) &&
1605 "Expected one VPActiveLaneMaskPHIRecipe for each unroll part");
1606
1607 auto *EntryALM = cast<VPInstruction>(Phis[0]->getStartValue());
1608 auto *LoopALM = cast<VPInstruction>(Phis[0]->getBackedgeValue());
1609
1610 assert((EntryALM->getOpcode() == VPInstruction::ActiveLaneMask &&
1611 LoopALM->getOpcode() == VPInstruction::ActiveLaneMask) &&
1612 "Expected incoming values of Phi to be ActiveLaneMasks");
1613
1614 // When using wide lane masks, the return type of the get.active.lane.mask
1615 // intrinsic is VF x UF (last operand).
1616 VPValue *ALMMultiplier =
1617 Plan.getOrAddLiveIn(ConstantInt::get(IntegerType::getInt64Ty(Ctx), UF));
1618 EntryALM->setOperand(2, ALMMultiplier);
1619 LoopALM->setOperand(2, ALMMultiplier);
1620
1621 // Create UF x extract vectors and insert into preheader.
1622 SmallVector<VPValue *> EntryExtracts(UF);
1623 ExtractFromALM(EntryALM, EntryExtracts);
1624
1625 // Create UF x extract vectors and insert before the loop compare & branch,
1626 // updating the compare to use the first extract.
1627 SmallVector<VPValue *> LoopExtracts(UF);
1628 ExtractFromALM(LoopALM, LoopExtracts);
1629 VPInstruction *Not = cast<VPInstruction>(Term->getOperand(0));
1630 Not->setOperand(0, LoopExtracts[0]);
1631
1632 // Update the incoming values of active lane mask phis.
1633 for (unsigned Part = 0; Part < UF; ++Part) {
1634 Phis[Part]->setStartValue(EntryExtracts[Part]);
1635 Phis[Part]->setBackedgeValue(LoopExtracts[Part]);
1636 }
1637
1638 return true;
1639}
1640
1641/// Try to simplify the branch condition of \p Plan. This may restrict the
1642/// resulting plan to \p BestVF and \p BestUF.
1644 unsigned BestUF,
1646 VPRegionBlock *VectorRegion = Plan.getVectorLoopRegion();
1647 VPBasicBlock *ExitingVPBB = VectorRegion->getExitingBasicBlock();
1648 auto *Term = &ExitingVPBB->back();
1649 VPValue *Cond;
1650 ScalarEvolution &SE = *PSE.getSE();
1651 if (match(Term, m_BranchOnCount(m_VPValue(), m_VPValue())) ||
1653 m_VPValue(), m_VPValue(), m_VPValue()))))) {
1654 // Try to simplify the branch condition if TC <= VF * UF when the latch
1655 // terminator is BranchOnCount or BranchOnCond where the input is
1656 // Not(ActiveLaneMask).
1657 const SCEV *TripCount =
1659 assert(!isa<SCEVCouldNotCompute>(TripCount) &&
1660 "Trip count SCEV must be computable");
1661 ElementCount NumElements = BestVF.multiplyCoefficientBy(BestUF);
1662 const SCEV *C = SE.getElementCount(TripCount->getType(), NumElements);
1663 if (TripCount->isZero() ||
1664 !SE.isKnownPredicate(CmpInst::ICMP_ULE, TripCount, C))
1665 return false;
1666 } else if (match(Term, m_BranchOnCond(m_VPValue(Cond)))) {
1667 // For BranchOnCond, check if we can prove the condition to be true using VF
1668 // and UF.
1669 if (!isConditionTrueViaVFAndUF(Cond, Plan, BestVF, BestUF, SE))
1670 return false;
1671 } else {
1672 return false;
1673 }
1674
1675 // The vector loop region only executes once. If possible, completely remove
1676 // the region, otherwise replace the terminator controlling the latch with
1677 // (BranchOnCond true).
1678 // TODO: VPWidenIntOrFpInductionRecipe is only partially supported; add
1679 // support for other non-canonical widen induction recipes (e.g.,
1680 // VPWidenPointerInductionRecipe).
1681 auto *Header = cast<VPBasicBlock>(VectorRegion->getEntry());
1682 if (all_of(Header->phis(), [](VPRecipeBase &Phi) {
1683 if (auto *R = dyn_cast<VPWidenIntOrFpInductionRecipe>(&Phi))
1684 return R->isCanonical();
1685 return isa<VPCanonicalIVPHIRecipe, VPEVLBasedIVPHIRecipe,
1686 VPFirstOrderRecurrencePHIRecipe, VPPhi>(&Phi);
1687 })) {
1688 for (VPRecipeBase &HeaderR : make_early_inc_range(Header->phis())) {
1689 if (auto *R = dyn_cast<VPWidenIntOrFpInductionRecipe>(&HeaderR)) {
1690 VPBuilder Builder(Plan.getVectorPreheader());
1691 VPValue *StepV = Builder.createNaryOp(VPInstruction::StepVector, {},
1692 R->getScalarType());
1693 HeaderR.getVPSingleValue()->replaceAllUsesWith(StepV);
1694 HeaderR.eraseFromParent();
1695 continue;
1696 }
1697 auto *Phi = cast<VPPhiAccessors>(&HeaderR);
1698 HeaderR.getVPSingleValue()->replaceAllUsesWith(Phi->getIncomingValue(0));
1699 HeaderR.eraseFromParent();
1700 }
1701
1702 VPBlockBase *Preheader = VectorRegion->getSinglePredecessor();
1703 VPBlockBase *Exit = VectorRegion->getSingleSuccessor();
1704 VPBlockUtils::disconnectBlocks(Preheader, VectorRegion);
1705 VPBlockUtils::disconnectBlocks(VectorRegion, Exit);
1706
1707 for (VPBlockBase *B : vp_depth_first_shallow(VectorRegion->getEntry()))
1708 B->setParent(nullptr);
1709
1710 VPBlockUtils::connectBlocks(Preheader, Header);
1711 VPBlockUtils::connectBlocks(ExitingVPBB, Exit);
1713 } else {
1714 // The vector region contains header phis for which we cannot remove the
1715 // loop region yet.
1716 auto *BOC = new VPInstruction(VPInstruction::BranchOnCond, {Plan.getTrue()},
1717 Term->getDebugLoc());
1718 ExitingVPBB->appendRecipe(BOC);
1719 }
1720
1721 Term->eraseFromParent();
1722
1723 return true;
1724}
1725
1727 unsigned BestUF,
1729 assert(Plan.hasVF(BestVF) && "BestVF is not available in Plan");
1730 assert(Plan.hasUF(BestUF) && "BestUF is not available in Plan");
1731
1732 bool MadeChange = tryToReplaceALMWithWideALM(Plan, BestVF, BestUF);
1733 MadeChange |= simplifyBranchConditionForVFAndUF(Plan, BestVF, BestUF, PSE);
1734 MadeChange |= optimizeVectorInductionWidthForTCAndVFUF(Plan, BestVF, BestUF);
1735
1736 if (MadeChange) {
1737 Plan.setVF(BestVF);
1738 assert(Plan.getUF() == BestUF && "BestUF must match the Plan's UF");
1739 }
1740 // TODO: Further simplifications are possible
1741 // 1. Replace inductions with constants.
1742 // 2. Replace vector loop region with VPBasicBlock.
1743}
1744
1745/// Sink users of \p FOR after the recipe defining the previous value \p
1746/// Previous of the recurrence. \returns true if all users of \p FOR could be
1747/// re-arranged as needed or false if it is not possible.
1748static bool
1750 VPRecipeBase *Previous,
1751 VPDominatorTree &VPDT) {
1752 // Collect recipes that need sinking.
1755 Seen.insert(Previous);
1756 auto TryToPushSinkCandidate = [&](VPRecipeBase *SinkCandidate) {
1757 // The previous value must not depend on the users of the recurrence phi. In
1758 // that case, FOR is not a fixed order recurrence.
1759 if (SinkCandidate == Previous)
1760 return false;
1761
1762 if (isa<VPHeaderPHIRecipe>(SinkCandidate) ||
1763 !Seen.insert(SinkCandidate).second ||
1764 VPDT.properlyDominates(Previous, SinkCandidate))
1765 return true;
1766
1767 if (SinkCandidate->mayHaveSideEffects())
1768 return false;
1769
1770 WorkList.push_back(SinkCandidate);
1771 return true;
1772 };
1773
1774 // Recursively sink users of FOR after Previous.
1775 WorkList.push_back(FOR);
1776 for (unsigned I = 0; I != WorkList.size(); ++I) {
1777 VPRecipeBase *Current = WorkList[I];
1778 assert(Current->getNumDefinedValues() == 1 &&
1779 "only recipes with a single defined value expected");
1780
1781 for (VPUser *User : Current->getVPSingleValue()->users()) {
1782 if (!TryToPushSinkCandidate(cast<VPRecipeBase>(User)))
1783 return false;
1784 }
1785 }
1786
1787 // Keep recipes to sink ordered by dominance so earlier instructions are
1788 // processed first.
1789 sort(WorkList, [&VPDT](const VPRecipeBase *A, const VPRecipeBase *B) {
1790 return VPDT.properlyDominates(A, B);
1791 });
1792
1793 for (VPRecipeBase *SinkCandidate : WorkList) {
1794 if (SinkCandidate == FOR)
1795 continue;
1796
1797 SinkCandidate->moveAfter(Previous);
1798 Previous = SinkCandidate;
1799 }
1800 return true;
1801}
1802
1803/// Try to hoist \p Previous and its operands before all users of \p FOR.
1805 VPRecipeBase *Previous,
1806 VPDominatorTree &VPDT) {
1807 if (Previous->mayHaveSideEffects() || Previous->mayReadFromMemory())
1808 return false;
1809
1810 // Collect recipes that need hoisting.
1811 SmallVector<VPRecipeBase *> HoistCandidates;
1813 VPRecipeBase *HoistPoint = nullptr;
1814 // Find the closest hoist point by looking at all users of FOR and selecting
1815 // the recipe dominating all other users.
1816 for (VPUser *U : FOR->users()) {
1817 auto *R = cast<VPRecipeBase>(U);
1818 if (!HoistPoint || VPDT.properlyDominates(R, HoistPoint))
1819 HoistPoint = R;
1820 }
1821 assert(all_of(FOR->users(),
1822 [&VPDT, HoistPoint](VPUser *U) {
1823 auto *R = cast<VPRecipeBase>(U);
1824 return HoistPoint == R ||
1825 VPDT.properlyDominates(HoistPoint, R);
1826 }) &&
1827 "HoistPoint must dominate all users of FOR");
1828
1829 auto NeedsHoisting = [HoistPoint, &VPDT,
1830 &Visited](VPValue *HoistCandidateV) -> VPRecipeBase * {
1831 VPRecipeBase *HoistCandidate = HoistCandidateV->getDefiningRecipe();
1832 if (!HoistCandidate)
1833 return nullptr;
1834 VPRegionBlock *EnclosingLoopRegion =
1835 HoistCandidate->getParent()->getEnclosingLoopRegion();
1836 assert((!HoistCandidate->getParent()->getParent() ||
1837 HoistCandidate->getParent()->getParent() == EnclosingLoopRegion) &&
1838 "CFG in VPlan should still be flat, without replicate regions");
1839 // Hoist candidate was already visited, no need to hoist.
1840 if (!Visited.insert(HoistCandidate).second)
1841 return nullptr;
1842
1843 // Candidate is outside loop region or a header phi, dominates FOR users w/o
1844 // hoisting.
1845 if (!EnclosingLoopRegion || isa<VPHeaderPHIRecipe>(HoistCandidate))
1846 return nullptr;
1847
1848 // If we reached a recipe that dominates HoistPoint, we don't need to
1849 // hoist the recipe.
1850 if (VPDT.properlyDominates(HoistCandidate, HoistPoint))
1851 return nullptr;
1852 return HoistCandidate;
1853 };
1854 auto CanHoist = [&](VPRecipeBase *HoistCandidate) {
1855 // Avoid hoisting candidates with side-effects, as we do not yet analyze
1856 // associated dependencies.
1857 return !HoistCandidate->mayHaveSideEffects();
1858 };
1859
1860 if (!NeedsHoisting(Previous->getVPSingleValue()))
1861 return true;
1862
1863 // Recursively try to hoist Previous and its operands before all users of FOR.
1864 HoistCandidates.push_back(Previous);
1865
1866 for (unsigned I = 0; I != HoistCandidates.size(); ++I) {
1867 VPRecipeBase *Current = HoistCandidates[I];
1868 assert(Current->getNumDefinedValues() == 1 &&
1869 "only recipes with a single defined value expected");
1870 if (!CanHoist(Current))
1871 return false;
1872
1873 for (VPValue *Op : Current->operands()) {
1874 // If we reach FOR, it means the original Previous depends on some other
1875 // recurrence that in turn depends on FOR. If that is the case, we would
1876 // also need to hoist recipes involving the other FOR, which may break
1877 // dependencies.
1878 if (Op == FOR)
1879 return false;
1880
1881 if (auto *R = NeedsHoisting(Op))
1882 HoistCandidates.push_back(R);
1883 }
1884 }
1885
1886 // Order recipes to hoist by dominance so earlier instructions are processed
1887 // first.
1888 sort(HoistCandidates, [&VPDT](const VPRecipeBase *A, const VPRecipeBase *B) {
1889 return VPDT.properlyDominates(A, B);
1890 });
1891
1892 for (VPRecipeBase *HoistCandidate : HoistCandidates) {
1893 HoistCandidate->moveBefore(*HoistPoint->getParent(),
1894 HoistPoint->getIterator());
1895 }
1896
1897 return true;
1898}
1899
1901 VPBuilder &LoopBuilder) {
1902 VPDominatorTree VPDT;
1903 VPDT.recalculate(Plan);
1904
1906 for (VPRecipeBase &R :
1909 RecurrencePhis.push_back(FOR);
1910
1911 for (VPFirstOrderRecurrencePHIRecipe *FOR : RecurrencePhis) {
1913 VPRecipeBase *Previous = FOR->getBackedgeValue()->getDefiningRecipe();
1914 // Fixed-order recurrences do not contain cycles, so this loop is guaranteed
1915 // to terminate.
1916 while (auto *PrevPhi =
1918 assert(PrevPhi->getParent() == FOR->getParent());
1919 assert(SeenPhis.insert(PrevPhi).second);
1920 Previous = PrevPhi->getBackedgeValue()->getDefiningRecipe();
1921 }
1922
1923 if (!sinkRecurrenceUsersAfterPrevious(FOR, Previous, VPDT) &&
1924 !hoistPreviousBeforeFORUsers(FOR, Previous, VPDT))
1925 return false;
1926
1927 // Introduce a recipe to combine the incoming and previous values of a
1928 // fixed-order recurrence.
1929 VPBasicBlock *InsertBlock = Previous->getParent();
1930 if (isa<VPHeaderPHIRecipe>(Previous))
1931 LoopBuilder.setInsertPoint(InsertBlock, InsertBlock->getFirstNonPhi());
1932 else
1933 LoopBuilder.setInsertPoint(InsertBlock,
1934 std::next(Previous->getIterator()));
1935
1936 auto *RecurSplice =
1938 {FOR, FOR->getBackedgeValue()});
1939
1940 FOR->replaceAllUsesWith(RecurSplice);
1941 // Set the first operand of RecurSplice to FOR again, after replacing
1942 // all users.
1943 RecurSplice->setOperand(0, FOR);
1944 }
1945 return true;
1946}
1947
1949 for (VPRecipeBase &R :
1951 auto *PhiR = dyn_cast<VPReductionPHIRecipe>(&R);
1952 if (!PhiR)
1953 continue;
1954 RecurKind RK = PhiR->getRecurrenceKind();
1955 if (RK != RecurKind::Add && RK != RecurKind::Mul && RK != RecurKind::Sub &&
1957 continue;
1958
1959 for (VPUser *U : collectUsersRecursively(PhiR))
1960 if (auto *RecWithFlags = dyn_cast<VPRecipeWithIRFlags>(U)) {
1961 RecWithFlags->dropPoisonGeneratingFlags();
1962 }
1963 }
1964}
1965
1966namespace {
1967struct VPCSEDenseMapInfo : public DenseMapInfo<VPSingleDefRecipe *> {
1968 static bool isSentinel(const VPSingleDefRecipe *Def) {
1969 return Def == getEmptyKey() || Def == getTombstoneKey();
1970 }
1971
1972 /// Get any instruction opcode or intrinsic ID data embedded in recipe \p R.
1973 /// Returns an optional pair, where the first element indicates whether it is
1974 /// an intrinsic ID.
1975 static std::optional<std::pair<bool, unsigned>>
1976 getOpcodeOrIntrinsicID(const VPSingleDefRecipe *R) {
1977 return TypeSwitch<const VPSingleDefRecipe *,
1978 std::optional<std::pair<bool, unsigned>>>(R)
1981 [](auto *I) { return std::make_pair(false, I->getOpcode()); })
1982 .Case<VPWidenIntrinsicRecipe>([](auto *I) {
1983 return std::make_pair(true, I->getVectorIntrinsicID());
1984 })
1985 .Default([](auto *) { return std::nullopt; });
1986 }
1987
1988 /// If recipe \p R will lower to a GEP with a non-i8 source element type,
1989 /// return that source element type.
1990 static Type *getGEPSourceElementType(const VPSingleDefRecipe *R) {
1991 // All VPInstructions that lower to GEPs must have the i8 source element
1992 // type (as they are PtrAdds), so we omit it.
1993 return TypeSwitch<const VPSingleDefRecipe *, Type *>(R)
1994 .Case<VPReplicateRecipe, VPWidenGEPRecipe>([](auto *I) -> Type * {
1995 if (auto *GEP = dyn_cast<GetElementPtrInst>(I->getUnderlyingValue()))
1996 return GEP->getSourceElementType();
1997 return nullptr;
1998 })
1999 .Case<VPVectorPointerRecipe>(
2000 [](auto *I) { return I->getSourceElementType(); })
2001 .Default([](auto *) { return nullptr; });
2002 }
2003
2004 /// Returns true if recipe \p Def can be safely handed for CSE.
2005 static bool canHandle(const VPSingleDefRecipe *Def) {
2006 // We can extend the list of handled recipes in the future,
2007 // provided we account for the data embedded in them while checking for
2008 // equality or hashing. We assign VPVectorEndPointerRecipe the GEP opcode,
2009 // as it is essentially a GEP with different semantics.
2010 auto C = isa<VPVectorPointerRecipe>(Def)
2011 ? std::make_pair(false, Instruction::GetElementPtr)
2012 : getOpcodeOrIntrinsicID(Def);
2013
2014 // The issue with (Insert|Extract)Value is that the index of the
2015 // insert/extract is not a proper operand in LLVM IR, and hence also not in
2016 // VPlan.
2017 if (!C || (!C->first && (C->second == Instruction::InsertValue ||
2018 C->second == Instruction::ExtractValue)))
2019 return false;
2020
2021 // During CSE, we can only handle recipes that don't read from memory: if
2022 // they read from memory, there could be an intervening write to memory
2023 // before the next instance is CSE'd, leading to an incorrect result.
2024 return !Def->mayReadFromMemory();
2025 }
2026
2027 /// Hash the underlying data of \p Def.
2028 static unsigned getHashValue(const VPSingleDefRecipe *Def) {
2029 const VPlan *Plan = Def->getParent()->getPlan();
2030 VPTypeAnalysis TypeInfo(*Plan);
2031 hash_code Result = hash_combine(
2032 Def->getVPDefID(), getOpcodeOrIntrinsicID(Def),
2033 getGEPSourceElementType(Def), TypeInfo.inferScalarType(Def),
2035 if (auto *RFlags = dyn_cast<VPRecipeWithIRFlags>(Def))
2036 if (RFlags->hasPredicate())
2037 return hash_combine(Result, RFlags->getPredicate());
2038 return Result;
2039 }
2040
2041 /// Check equality of underlying data of \p L and \p R.
2042 static bool isEqual(const VPSingleDefRecipe *L, const VPSingleDefRecipe *R) {
2043 if (isSentinel(L) || isSentinel(R))
2044 return L == R;
2045 if (L->getVPDefID() != R->getVPDefID() ||
2046 getOpcodeOrIntrinsicID(L) != getOpcodeOrIntrinsicID(R) ||
2047 getGEPSourceElementType(L) != getGEPSourceElementType(R) ||
2049 !equal(L->operands(), R->operands()))
2050 return false;
2051 if (auto *LFlags = dyn_cast<VPRecipeWithIRFlags>(L))
2052 if (LFlags->hasPredicate() &&
2053 LFlags->getPredicate() !=
2054 cast<VPRecipeWithIRFlags>(R)->getPredicate())
2055 return false;
2056 const VPlan *Plan = L->getParent()->getPlan();
2057 VPTypeAnalysis TypeInfo(*Plan);
2058 return TypeInfo.inferScalarType(L) == TypeInfo.inferScalarType(R);
2059 }
2060};
2061} // end anonymous namespace
2062
2063/// Perform a common-subexpression-elimination of VPSingleDefRecipes on the \p
2064/// Plan.
2066 VPDominatorTree VPDT(Plan);
2068
2070 vp_depth_first_deep(Plan.getEntry()))) {
2071 for (VPRecipeBase &R : *VPBB) {
2072 auto *Def = dyn_cast<VPSingleDefRecipe>(&R);
2073 if (!Def || !VPCSEDenseMapInfo::canHandle(Def))
2074 continue;
2075 if (VPSingleDefRecipe *V = CSEMap.lookup(Def)) {
2076 // V must dominate Def for a valid replacement.
2077 if (!VPDT.dominates(V->getParent(), VPBB))
2078 continue;
2079 // Only keep flags present on both V and Def.
2080 if (auto *RFlags = dyn_cast<VPRecipeWithIRFlags>(V))
2081 RFlags->intersectFlags(*cast<VPRecipeWithIRFlags>(Def));
2082 Def->replaceAllUsesWith(V);
2083 continue;
2084 }
2085 CSEMap[Def] = Def;
2086 }
2087 }
2088}
2089
2090/// Move loop-invariant recipes out of the vector loop region in \p Plan.
2091static void licm(VPlan &Plan) {
2092 VPBasicBlock *Preheader = Plan.getVectorPreheader();
2093
2094 // Return true if we do not know how to (mechanically) hoist a given recipe
2095 // out of a loop region. Does not address legality concerns such as aliasing
2096 // or speculation safety.
2097 auto CannotHoistRecipe = [](VPRecipeBase &R) {
2098 // Allocas cannot be hoisted.
2099 auto *RepR = dyn_cast<VPReplicateRecipe>(&R);
2100 return RepR && RepR->getOpcode() == Instruction::Alloca;
2101 };
2102
2103 // Hoist any loop invariant recipes from the vector loop region to the
2104 // preheader. Preform a shallow traversal of the vector loop region, to
2105 // exclude recipes in replicate regions.
2106 VPRegionBlock *LoopRegion = Plan.getVectorLoopRegion();
2108 vp_depth_first_shallow(LoopRegion->getEntry()))) {
2109 for (VPRecipeBase &R : make_early_inc_range(*VPBB)) {
2110 if (CannotHoistRecipe(R))
2111 continue;
2112 // TODO: Relax checks in the future, e.g. we could also hoist reads, if
2113 // their memory location is not modified in the vector loop.
2114 if (R.mayHaveSideEffects() || R.mayReadFromMemory() || R.isPhi() ||
2115 any_of(R.operands(), [](VPValue *Op) {
2116 return !Op->isDefinedOutsideLoopRegions();
2117 }))
2118 continue;
2119 R.moveBefore(*Preheader, Preheader->end());
2120 }
2121 }
2122}
2123
2125 VPlan &Plan, const MapVector<Instruction *, uint64_t> &MinBWs) {
2126 // Keep track of created truncates, so they can be re-used. Note that we
2127 // cannot use RAUW after creating a new truncate, as this would could make
2128 // other uses have different types for their operands, making them invalidly
2129 // typed.
2131 VPTypeAnalysis TypeInfo(Plan);
2132 VPBasicBlock *PH = Plan.getVectorPreheader();
2135 for (VPRecipeBase &R : make_early_inc_range(*VPBB)) {
2138 &R))
2139 continue;
2140
2141 VPValue *ResultVPV = R.getVPSingleValue();
2142 auto *UI = cast_or_null<Instruction>(ResultVPV->getUnderlyingValue());
2143 unsigned NewResSizeInBits = MinBWs.lookup(UI);
2144 if (!NewResSizeInBits)
2145 continue;
2146
2147 // If the value wasn't vectorized, we must maintain the original scalar
2148 // type. Skip those here, after incrementing NumProcessedRecipes. Also
2149 // skip casts which do not need to be handled explicitly here, as
2150 // redundant casts will be removed during recipe simplification.
2152 continue;
2153
2154 Type *OldResTy = TypeInfo.inferScalarType(ResultVPV);
2155 unsigned OldResSizeInBits = OldResTy->getScalarSizeInBits();
2156 assert(OldResTy->isIntegerTy() && "only integer types supported");
2157 (void)OldResSizeInBits;
2158
2159 auto *NewResTy = IntegerType::get(Plan.getContext(), NewResSizeInBits);
2160
2161 // Any wrapping introduced by shrinking this operation shouldn't be
2162 // considered undefined behavior. So, we can't unconditionally copy
2163 // arithmetic wrapping flags to VPW.
2164 if (auto *VPW = dyn_cast<VPRecipeWithIRFlags>(&R))
2165 VPW->dropPoisonGeneratingFlags();
2166
2167 if (OldResSizeInBits != NewResSizeInBits &&
2168 !match(&R, m_ICmp(m_VPValue(), m_VPValue()))) {
2169 // Extend result to original width.
2170 auto *Ext =
2171 new VPWidenCastRecipe(Instruction::ZExt, ResultVPV, OldResTy);
2172 Ext->insertAfter(&R);
2173 ResultVPV->replaceAllUsesWith(Ext);
2174 Ext->setOperand(0, ResultVPV);
2175 assert(OldResSizeInBits > NewResSizeInBits && "Nothing to shrink?");
2176 } else {
2177 assert(match(&R, m_ICmp(m_VPValue(), m_VPValue())) &&
2178 "Only ICmps should not need extending the result.");
2179 }
2180
2181 assert(!isa<VPWidenStoreRecipe>(&R) && "stores cannot be narrowed");
2183 continue;
2184
2185 // Shrink operands by introducing truncates as needed.
2186 unsigned StartIdx = isa<VPWidenSelectRecipe>(&R) ? 1 : 0;
2187 for (unsigned Idx = StartIdx; Idx != R.getNumOperands(); ++Idx) {
2188 auto *Op = R.getOperand(Idx);
2189 unsigned OpSizeInBits =
2191 if (OpSizeInBits == NewResSizeInBits)
2192 continue;
2193 assert(OpSizeInBits > NewResSizeInBits && "nothing to truncate");
2194 auto [ProcessedIter, IterIsEmpty] = ProcessedTruncs.try_emplace(Op);
2195 VPWidenCastRecipe *NewOp =
2196 IterIsEmpty
2197 ? new VPWidenCastRecipe(Instruction::Trunc, Op, NewResTy)
2198 : ProcessedIter->second;
2199 R.setOperand(Idx, NewOp);
2200 if (!IterIsEmpty)
2201 continue;
2202 ProcessedIter->second = NewOp;
2203 if (!Op->isLiveIn()) {
2204 NewOp->insertBefore(&R);
2205 } else {
2206 PH->appendRecipe(NewOp);
2207 }
2208 }
2209
2210 }
2211 }
2212}
2213
2217 VPValue *Cond;
2218 // Skip blocks that are not terminated by BranchOnCond.
2219 if (VPBB->empty() || !match(&VPBB->back(), m_BranchOnCond(m_VPValue(Cond))))
2220 continue;
2221
2222 assert(VPBB->getNumSuccessors() == 2 &&
2223 "Two successors expected for BranchOnCond");
2224 unsigned RemovedIdx;
2225 if (match(Cond, m_True()))
2226 RemovedIdx = 1;
2227 else if (match(Cond, m_False()))
2228 RemovedIdx = 0;
2229 else
2230 continue;
2231
2232 VPBasicBlock *RemovedSucc =
2233 cast<VPBasicBlock>(VPBB->getSuccessors()[RemovedIdx]);
2234 assert(count(RemovedSucc->getPredecessors(), VPBB) == 1 &&
2235 "There must be a single edge between VPBB and its successor");
2236 // Values coming from VPBB into phi recipes of RemoveSucc are removed from
2237 // these recipes.
2238 for (VPRecipeBase &R : RemovedSucc->phis())
2239 cast<VPPhiAccessors>(&R)->removeIncomingValueFor(VPBB);
2240
2241 // Disconnect blocks and remove the terminator. RemovedSucc will be deleted
2242 // automatically on VPlan destruction if it becomes unreachable.
2243 VPBlockUtils::disconnectBlocks(VPBB, RemovedSucc);
2244 VPBB->back().eraseFromParent();
2245 }
2246}
2247
2266
2267// Add a VPActiveLaneMaskPHIRecipe and related recipes to \p Plan and replace
2268// the loop terminator with a branch-on-cond recipe with the negated
2269// active-lane-mask as operand. Note that this turns the loop into an
2270// uncountable one. Only the existing terminator is replaced, all other existing
2271// recipes/users remain unchanged, except for poison-generating flags being
2272// dropped from the canonical IV increment. Return the created
2273// VPActiveLaneMaskPHIRecipe.
2274//
2275// The function uses the following definitions:
2276//
2277// %TripCount = DataWithControlFlowWithoutRuntimeCheck ?
2278// calculate-trip-count-minus-VF (original TC) : original TC
2279// %IncrementValue = DataWithControlFlowWithoutRuntimeCheck ?
2280// CanonicalIVPhi : CanonicalIVIncrement
2281// %StartV is the canonical induction start value.
2282//
2283// The function adds the following recipes:
2284//
2285// vector.ph:
2286// %TripCount = calculate-trip-count-minus-VF (original TC)
2287// [if DataWithControlFlowWithoutRuntimeCheck]
2288// %EntryInc = canonical-iv-increment-for-part %StartV
2289// %EntryALM = active-lane-mask %EntryInc, %TripCount
2290//
2291// vector.body:
2292// ...
2293// %P = active-lane-mask-phi [ %EntryALM, %vector.ph ], [ %ALM, %vector.body ]
2294// ...
2295// %InLoopInc = canonical-iv-increment-for-part %IncrementValue
2296// %ALM = active-lane-mask %InLoopInc, TripCount
2297// %Negated = Not %ALM
2298// branch-on-cond %Negated
2299//
2302 VPRegionBlock *TopRegion = Plan.getVectorLoopRegion();
2303 VPBasicBlock *EB = TopRegion->getExitingBasicBlock();
2304 auto *CanonicalIVPHI = Plan.getCanonicalIV();
2305 VPValue *StartV = CanonicalIVPHI->getStartValue();
2306
2307 auto *CanonicalIVIncrement =
2308 cast<VPInstruction>(CanonicalIVPHI->getBackedgeValue());
2309 // TODO: Check if dropping the flags is needed if
2310 // !DataAndControlFlowWithoutRuntimeCheck.
2311 CanonicalIVIncrement->dropPoisonGeneratingFlags();
2312 DebugLoc DL = CanonicalIVIncrement->getDebugLoc();
2313 // We can't use StartV directly in the ActiveLaneMask VPInstruction, since
2314 // we have to take unrolling into account. Each part needs to start at
2315 // Part * VF
2316 auto *VecPreheader = Plan.getVectorPreheader();
2317 VPBuilder Builder(VecPreheader);
2318
2319 // Create the ActiveLaneMask instruction using the correct start values.
2320 VPValue *TC = Plan.getTripCount();
2321
2322 VPValue *TripCount, *IncrementValue;
2324 // When the loop is guarded by a runtime overflow check for the loop
2325 // induction variable increment by VF, we can increment the value before
2326 // the get.active.lane mask and use the unmodified tripcount.
2327 IncrementValue = CanonicalIVIncrement;
2328 TripCount = TC;
2329 } else {
2330 // When avoiding a runtime check, the active.lane.mask inside the loop
2331 // uses a modified trip count and the induction variable increment is
2332 // done after the active.lane.mask intrinsic is called.
2333 IncrementValue = CanonicalIVPHI;
2334 TripCount = Builder.createNaryOp(VPInstruction::CalculateTripCountMinusVF,
2335 {TC}, DL);
2336 }
2337 auto *EntryIncrement = Builder.createOverflowingOp(
2338 VPInstruction::CanonicalIVIncrementForPart, {StartV}, {false, false}, DL,
2339 "index.part.next");
2340
2341 // Create the active lane mask instruction in the VPlan preheader.
2342 VPValue *ALMMultiplier = Plan.getOrAddLiveIn(
2343 ConstantInt::get(Plan.getCanonicalIV()->getScalarType(), 1));
2344 auto *EntryALM = Builder.createNaryOp(VPInstruction::ActiveLaneMask,
2345 {EntryIncrement, TC, ALMMultiplier}, DL,
2346 "active.lane.mask.entry");
2347
2348 // Now create the ActiveLaneMaskPhi recipe in the main loop using the
2349 // preheader ActiveLaneMask instruction.
2350 auto *LaneMaskPhi =
2352 LaneMaskPhi->insertAfter(CanonicalIVPHI);
2353
2354 // Create the active lane mask for the next iteration of the loop before the
2355 // original terminator.
2356 VPRecipeBase *OriginalTerminator = EB->getTerminator();
2357 Builder.setInsertPoint(OriginalTerminator);
2358 auto *InLoopIncrement =
2359 Builder.createOverflowingOp(VPInstruction::CanonicalIVIncrementForPart,
2360 {IncrementValue}, {false, false}, DL);
2361 auto *ALM = Builder.createNaryOp(VPInstruction::ActiveLaneMask,
2362 {InLoopIncrement, TripCount, ALMMultiplier},
2363 DL, "active.lane.mask.next");
2364 LaneMaskPhi->addOperand(ALM);
2365
2366 // Replace the original terminator with BranchOnCond. We have to invert the
2367 // mask here because a true condition means jumping to the exit block.
2368 auto *NotMask = Builder.createNot(ALM, DL);
2369 Builder.createNaryOp(VPInstruction::BranchOnCond, {NotMask}, DL);
2370 OriginalTerminator->eraseFromParent();
2371 return LaneMaskPhi;
2372}
2373
2374/// Collect the header mask with the pattern:
2375/// (ICMP_ULE, WideCanonicalIV, backedge-taken-count)
2376/// TODO: Introduce explicit recipe for header-mask instead of searching
2377/// for the header-mask pattern manually.
2379 SmallVector<VPValue *> WideCanonicalIVs;
2380 auto *FoundWidenCanonicalIVUser = find_if(Plan.getCanonicalIV()->users(),
2384 "Must have at most one VPWideCanonicalIVRecipe");
2385 if (FoundWidenCanonicalIVUser != Plan.getCanonicalIV()->users().end()) {
2386 auto *WideCanonicalIV =
2387 cast<VPWidenCanonicalIVRecipe>(*FoundWidenCanonicalIVUser);
2388 WideCanonicalIVs.push_back(WideCanonicalIV);
2389 }
2390
2391 // Also include VPWidenIntOrFpInductionRecipes that represent a widened
2392 // version of the canonical induction.
2393 VPBasicBlock *HeaderVPBB = Plan.getVectorLoopRegion()->getEntryBasicBlock();
2394 for (VPRecipeBase &Phi : HeaderVPBB->phis()) {
2395 auto *WidenOriginalIV = dyn_cast<VPWidenIntOrFpInductionRecipe>(&Phi);
2396 if (WidenOriginalIV && WidenOriginalIV->isCanonical())
2397 WideCanonicalIVs.push_back(WidenOriginalIV);
2398 }
2399
2400 // Walk users of wide canonical IVs and find the single compare of the form
2401 // (ICMP_ULE, WideCanonicalIV, backedge-taken-count).
2402 VPSingleDefRecipe *HeaderMask = nullptr;
2403 for (auto *Wide : WideCanonicalIVs) {
2404 for (VPUser *U : SmallVector<VPUser *>(Wide->users())) {
2405 auto *VPI = dyn_cast<VPInstruction>(U);
2406 if (!VPI || !vputils::isHeaderMask(VPI, Plan))
2407 continue;
2408
2409 assert(VPI->getOperand(0) == Wide &&
2410 "WidenCanonicalIV must be the first operand of the compare");
2411 assert(!HeaderMask && "Multiple header masks found?");
2412 HeaderMask = VPI;
2413 }
2414 }
2415 return HeaderMask;
2416}
2417
2419 VPlan &Plan, bool UseActiveLaneMaskForControlFlow,
2422 UseActiveLaneMaskForControlFlow) &&
2423 "DataAndControlFlowWithoutRuntimeCheck implies "
2424 "UseActiveLaneMaskForControlFlow");
2425
2426 auto *FoundWidenCanonicalIVUser = find_if(Plan.getCanonicalIV()->users(),
2428 assert(FoundWidenCanonicalIVUser &&
2429 "Must have widened canonical IV when tail folding!");
2430 VPSingleDefRecipe *HeaderMask = findHeaderMask(Plan);
2431 auto *WideCanonicalIV =
2432 cast<VPWidenCanonicalIVRecipe>(*FoundWidenCanonicalIVUser);
2433 VPSingleDefRecipe *LaneMask;
2434 if (UseActiveLaneMaskForControlFlow) {
2437 } else {
2438 VPBuilder B = VPBuilder::getToInsertAfter(WideCanonicalIV);
2439 VPValue *ALMMultiplier = Plan.getOrAddLiveIn(
2440 ConstantInt::get(Plan.getCanonicalIV()->getScalarType(), 1));
2441 LaneMask =
2442 B.createNaryOp(VPInstruction::ActiveLaneMask,
2443 {WideCanonicalIV, Plan.getTripCount(), ALMMultiplier},
2444 nullptr, "active.lane.mask");
2445 }
2446
2447 // Walk users of WideCanonicalIV and replace the header mask of the form
2448 // (ICMP_ULE, WideCanonicalIV, backedge-taken-count) with an active-lane-mask,
2449 // removing the old one to ensure there is always only a single header mask.
2450 HeaderMask->replaceAllUsesWith(LaneMask);
2451 HeaderMask->eraseFromParent();
2452}
2453
2454/// Try to optimize a \p CurRecipe masked by \p HeaderMask to a corresponding
2455/// EVL-based recipe without the header mask. Returns nullptr if no EVL-based
2456/// recipe could be created.
2457/// \p HeaderMask Header Mask.
2458/// \p CurRecipe Recipe to be transform.
2459/// \p TypeInfo VPlan-based type analysis.
2460/// \p AllOneMask The vector mask parameter of vector-predication intrinsics.
2461/// \p EVL The explicit vector length parameter of vector-predication
2462/// intrinsics.
2464 VPRecipeBase &CurRecipe,
2465 VPTypeAnalysis &TypeInfo,
2466 VPValue &AllOneMask, VPValue &EVL) {
2467 // FIXME: Don't transform recipes to EVL recipes if they're not masked by the
2468 // header mask.
2469 auto GetNewMask = [&](VPValue *OrigMask) -> VPValue * {
2470 assert(OrigMask && "Unmasked recipe when folding tail");
2471 // HeaderMask will be handled using EVL.
2472 VPValue *Mask;
2473 if (match(OrigMask, m_LogicalAnd(m_Specific(HeaderMask), m_VPValue(Mask))))
2474 return Mask;
2475 return HeaderMask == OrigMask ? nullptr : OrigMask;
2476 };
2477
2478 /// Adjust any end pointers so that they point to the end of EVL lanes not VF.
2479 auto GetNewAddr = [&CurRecipe, &EVL](VPValue *Addr) -> VPValue * {
2480 auto *EndPtr = dyn_cast<VPVectorEndPointerRecipe>(Addr);
2481 if (!EndPtr)
2482 return Addr;
2483 assert(EndPtr->getOperand(1) == &EndPtr->getParent()->getPlan()->getVF() &&
2484 "VPVectorEndPointerRecipe with non-VF VF operand?");
2485 assert(
2486 all_of(EndPtr->users(),
2487 [](VPUser *U) {
2488 return cast<VPWidenMemoryRecipe>(U)->isReverse();
2489 }) &&
2490 "VPVectorEndPointRecipe not used by reversed widened memory recipe?");
2491 VPVectorEndPointerRecipe *EVLAddr = EndPtr->clone();
2492 EVLAddr->insertBefore(&CurRecipe);
2493 EVLAddr->setOperand(1, &EVL);
2494 return EVLAddr;
2495 };
2496
2499 VPValue *NewMask = GetNewMask(L->getMask());
2500 VPValue *NewAddr = GetNewAddr(L->getAddr());
2501 return new VPWidenLoadEVLRecipe(*L, NewAddr, EVL, NewMask);
2502 })
2503 .Case<VPWidenStoreRecipe>([&](VPWidenStoreRecipe *S) {
2504 VPValue *NewMask = GetNewMask(S->getMask());
2505 VPValue *NewAddr = GetNewAddr(S->getAddr());
2506 return new VPWidenStoreEVLRecipe(*S, NewAddr, EVL, NewMask);
2507 })
2508 .Case<VPInterleaveRecipe>([&](VPInterleaveRecipe *IR) {
2509 VPValue *NewMask = GetNewMask(IR->getMask());
2510 return new VPInterleaveEVLRecipe(*IR, EVL, NewMask);
2511 })
2512 .Case<VPReductionRecipe>([&](VPReductionRecipe *Red) {
2513 VPValue *NewMask = GetNewMask(Red->getCondOp());
2514 return new VPReductionEVLRecipe(*Red, EVL, NewMask);
2515 })
2516 .Case<VPInstruction>([&](VPInstruction *VPI) -> VPRecipeBase * {
2517 VPValue *LHS, *RHS;
2518 // Transform select with a header mask condition
2519 // select(header_mask, LHS, RHS)
2520 // into vector predication merge.
2521 // vp.merge(all-true, LHS, RHS, EVL)
2522 if (!match(VPI, m_Select(m_Specific(HeaderMask), m_VPValue(LHS),
2523 m_VPValue(RHS))))
2524 return nullptr;
2525 // Use all true as the condition because this transformation is
2526 // limited to selects whose condition is a header mask.
2527 return new VPWidenIntrinsicRecipe(
2528 Intrinsic::vp_merge, {&AllOneMask, LHS, RHS, &EVL},
2529 TypeInfo.inferScalarType(LHS), VPI->getDebugLoc());
2530 })
2531 .Default([&](VPRecipeBase *R) { return nullptr; });
2532}
2533
2534/// Replace recipes with their EVL variants.
2536 VPTypeAnalysis TypeInfo(Plan);
2537 VPValue *AllOneMask = Plan.getTrue();
2538 VPRegionBlock *LoopRegion = Plan.getVectorLoopRegion();
2539 VPBasicBlock *Header = LoopRegion->getEntryBasicBlock();
2540
2541 assert(all_of(Plan.getVF().users(),
2544 "User of VF that we can't transform to EVL.");
2545 Plan.getVF().replaceUsesWithIf(&EVL, [](VPUser &U, unsigned Idx) {
2547 });
2548
2549 assert(all_of(Plan.getVFxUF().users(),
2550 [&Plan](VPUser *U) {
2551 return match(U, m_c_Add(m_Specific(Plan.getCanonicalIV()),
2552 m_Specific(&Plan.getVFxUF()))) ||
2553 isa<VPWidenPointerInductionRecipe>(U);
2554 }) &&
2555 "Only users of VFxUF should be VPWidenPointerInductionRecipe and the "
2556 "increment of the canonical induction.");
2557 Plan.getVFxUF().replaceUsesWithIf(&EVL, [](VPUser &U, unsigned Idx) {
2558 // Only replace uses in VPWidenPointerInductionRecipe; The increment of the
2559 // canonical induction must not be updated.
2561 });
2562
2563 // Defer erasing recipes till the end so that we don't invalidate the
2564 // VPTypeAnalysis cache.
2566
2567 // Create a scalar phi to track the previous EVL if fixed-order recurrence is
2568 // contained.
2569 bool ContainsFORs =
2571 if (ContainsFORs) {
2572 // TODO: Use VPInstruction::ExplicitVectorLength to get maximum EVL.
2573 VPValue *MaxEVL = &Plan.getVF();
2574 // Emit VPScalarCastRecipe in preheader if VF is not a 32 bits integer.
2575 VPBuilder Builder(LoopRegion->getPreheaderVPBB());
2576 MaxEVL = Builder.createScalarZExtOrTrunc(
2577 MaxEVL, Type::getInt32Ty(Plan.getContext()),
2578 TypeInfo.inferScalarType(MaxEVL), DebugLoc::getUnknown());
2579
2580 Builder.setInsertPoint(Header, Header->getFirstNonPhi());
2581 VPValue *PrevEVL = Builder.createScalarPhi(
2582 {MaxEVL, &EVL}, DebugLoc::getUnknown(), "prev.evl");
2583
2586 for (VPRecipeBase &R : *VPBB) {
2587 VPValue *V1, *V2;
2588 if (!match(&R,
2590 m_VPValue(V1), m_VPValue(V2))))
2591 continue;
2592 VPValue *Imm = Plan.getOrAddLiveIn(
2595 Intrinsic::experimental_vp_splice,
2596 {V1, V2, Imm, AllOneMask, PrevEVL, &EVL},
2597 TypeInfo.inferScalarType(R.getVPSingleValue()), R.getDebugLoc());
2598 VPSplice->insertBefore(&R);
2599 R.getVPSingleValue()->replaceAllUsesWith(VPSplice);
2600 ToErase.push_back(&R);
2601 }
2602 }
2603 }
2604
2605 VPValue *HeaderMask = findHeaderMask(Plan);
2606 if (!HeaderMask)
2607 return;
2608
2609 // Replace header masks with a mask equivalent to predicating by EVL:
2610 //
2611 // icmp ule widen-canonical-iv backedge-taken-count
2612 // ->
2613 // icmp ult step-vector, EVL
2614 VPRecipeBase *EVLR = EVL.getDefiningRecipe();
2615 VPBuilder Builder(EVLR->getParent(), std::next(EVLR->getIterator()));
2616 Type *EVLType = TypeInfo.inferScalarType(&EVL);
2617 VPValue *EVLMask = Builder.createICmp(
2619 Builder.createNaryOp(VPInstruction::StepVector, {}, EVLType), &EVL);
2620 HeaderMask->replaceAllUsesWith(EVLMask);
2621 ToErase.push_back(HeaderMask->getDefiningRecipe());
2622
2623 // Try to optimize header mask recipes away to their EVL variants.
2624 // TODO: Split optimizeMaskToEVL out and move into
2625 // VPlanTransforms::optimize. transformRecipestoEVLRecipes should be run in
2626 // tryToBuildVPlanWithVPRecipes beforehand.
2627 for (VPUser *U : collectUsersRecursively(EVLMask)) {
2628 auto *CurRecipe = cast<VPRecipeBase>(U);
2629 VPRecipeBase *EVLRecipe =
2630 optimizeMaskToEVL(EVLMask, *CurRecipe, TypeInfo, *AllOneMask, EVL);
2631 if (!EVLRecipe)
2632 continue;
2633
2634 unsigned NumDefVal = EVLRecipe->getNumDefinedValues();
2635 assert(NumDefVal == CurRecipe->getNumDefinedValues() &&
2636 "New recipe must define the same number of values as the "
2637 "original.");
2638 EVLRecipe->insertBefore(CurRecipe);
2640 EVLRecipe)) {
2641 for (unsigned I = 0; I < NumDefVal; ++I) {
2642 VPValue *CurVPV = CurRecipe->getVPValue(I);
2643 CurVPV->replaceAllUsesWith(EVLRecipe->getVPValue(I));
2644 }
2645 }
2646 ToErase.push_back(CurRecipe);
2647 }
2648 // Remove dead EVL mask.
2649 if (EVLMask->getNumUsers() == 0)
2650 ToErase.push_back(EVLMask->getDefiningRecipe());
2651
2652 for (VPRecipeBase *R : reverse(ToErase)) {
2653 SmallVector<VPValue *> PossiblyDead(R->operands());
2654 R->eraseFromParent();
2655 for (VPValue *Op : PossiblyDead)
2657 }
2658}
2659
2660/// Add a VPEVLBasedIVPHIRecipe and related recipes to \p Plan and
2661/// replaces all uses except the canonical IV increment of
2662/// VPCanonicalIVPHIRecipe with a VPEVLBasedIVPHIRecipe. VPCanonicalIVPHIRecipe
2663/// is used only for loop iterations counting after this transformation.
2664///
2665/// The function uses the following definitions:
2666/// %StartV is the canonical induction start value.
2667///
2668/// The function adds the following recipes:
2669///
2670/// vector.ph:
2671/// ...
2672///
2673/// vector.body:
2674/// ...
2675/// %EVLPhi = EXPLICIT-VECTOR-LENGTH-BASED-IV-PHI [ %StartV, %vector.ph ],
2676/// [ %NextEVLIV, %vector.body ]
2677/// %AVL = phi [ trip-count, %vector.ph ], [ %NextAVL, %vector.body ]
2678/// %VPEVL = EXPLICIT-VECTOR-LENGTH %AVL
2679/// ...
2680/// %OpEVL = cast i32 %VPEVL to IVSize
2681/// %NextEVLIV = add IVSize %OpEVL, %EVLPhi
2682/// %NextAVL = sub IVSize nuw %AVL, %OpEVL
2683/// ...
2684///
2685/// If MaxSafeElements is provided, the function adds the following recipes:
2686/// vector.ph:
2687/// ...
2688///
2689/// vector.body:
2690/// ...
2691/// %EVLPhi = EXPLICIT-VECTOR-LENGTH-BASED-IV-PHI [ %StartV, %vector.ph ],
2692/// [ %NextEVLIV, %vector.body ]
2693/// %AVL = phi [ trip-count, %vector.ph ], [ %NextAVL, %vector.body ]
2694/// %cmp = cmp ult %AVL, MaxSafeElements
2695/// %SAFE_AVL = select %cmp, %AVL, MaxSafeElements
2696/// %VPEVL = EXPLICIT-VECTOR-LENGTH %SAFE_AVL
2697/// ...
2698/// %OpEVL = cast i32 %VPEVL to IVSize
2699/// %NextEVLIV = add IVSize %OpEVL, %EVLPhi
2700/// %NextAVL = sub IVSize nuw %AVL, %OpEVL
2701/// ...
2702///
2704 VPlan &Plan, const std::optional<unsigned> &MaxSafeElements) {
2706
2707 auto *CanonicalIVPHI = Plan.getCanonicalIV();
2708 auto *CanIVTy = CanonicalIVPHI->getScalarType();
2709 VPValue *StartV = CanonicalIVPHI->getStartValue();
2710
2711 // Create the ExplicitVectorLengthPhi recipe in the main loop.
2712 auto *EVLPhi = new VPEVLBasedIVPHIRecipe(StartV, DebugLoc::getUnknown());
2713 EVLPhi->insertAfter(CanonicalIVPHI);
2714 VPBuilder Builder(Header, Header->getFirstNonPhi());
2715 // Create the AVL (application vector length), starting from TC -> 0 in steps
2716 // of EVL.
2717 VPPhi *AVLPhi = Builder.createScalarPhi(
2718 {Plan.getTripCount()}, DebugLoc::getCompilerGenerated(), "avl");
2719 VPValue *AVL = AVLPhi;
2720
2721 if (MaxSafeElements) {
2722 // Support for MaxSafeDist for correct loop emission.
2723 VPValue *AVLSafe =
2724 Plan.getOrAddLiveIn(ConstantInt::get(CanIVTy, *MaxSafeElements));
2725 VPValue *Cmp = Builder.createICmp(ICmpInst::ICMP_ULT, AVL, AVLSafe);
2726 AVL = Builder.createSelect(Cmp, AVL, AVLSafe, DebugLoc::getUnknown(),
2727 "safe_avl");
2728 }
2729 auto *VPEVL = Builder.createNaryOp(VPInstruction::ExplicitVectorLength, AVL,
2731
2732 auto *CanonicalIVIncrement =
2733 cast<VPInstruction>(CanonicalIVPHI->getBackedgeValue());
2734 Builder.setInsertPoint(CanonicalIVIncrement);
2735 VPValue *OpVPEVL = VPEVL;
2736
2737 auto *I32Ty = Type::getInt32Ty(Plan.getContext());
2738 OpVPEVL = Builder.createScalarZExtOrTrunc(
2739 OpVPEVL, CanIVTy, I32Ty, CanonicalIVIncrement->getDebugLoc());
2740
2741 auto *NextEVLIV = Builder.createOverflowingOp(
2742 Instruction::Add, {OpVPEVL, EVLPhi},
2743 {CanonicalIVIncrement->hasNoUnsignedWrap(),
2744 CanonicalIVIncrement->hasNoSignedWrap()},
2745 CanonicalIVIncrement->getDebugLoc(), "index.evl.next");
2746 EVLPhi->addOperand(NextEVLIV);
2747
2748 VPValue *NextAVL = Builder.createOverflowingOp(
2749 Instruction::Sub, {AVLPhi, OpVPEVL}, {/*hasNUW=*/true, /*hasNSW=*/false},
2750 DebugLoc::getCompilerGenerated(), "avl.next");
2751 AVLPhi->addOperand(NextAVL);
2752
2753 transformRecipestoEVLRecipes(Plan, *VPEVL);
2754
2755 // Replace all uses of VPCanonicalIVPHIRecipe by
2756 // VPEVLBasedIVPHIRecipe except for the canonical IV increment.
2757 CanonicalIVPHI->replaceAllUsesWith(EVLPhi);
2758 CanonicalIVIncrement->setOperand(0, CanonicalIVPHI);
2759 // TODO: support unroll factor > 1.
2760 Plan.setUF(1);
2761}
2762
2764 // Find EVL loop entries by locating VPEVLBasedIVPHIRecipe.
2765 // There should be only one EVL PHI in the entire plan.
2766 VPEVLBasedIVPHIRecipe *EVLPhi = nullptr;
2767
2770 for (VPRecipeBase &R : VPBB->phis())
2771 if (auto *PhiR = dyn_cast<VPEVLBasedIVPHIRecipe>(&R)) {
2772 assert(!EVLPhi && "Found multiple EVL PHIs. Only one expected");
2773 EVLPhi = PhiR;
2774 }
2775
2776 // Early return if no EVL PHI is found.
2777 if (!EVLPhi)
2778 return;
2779
2780 VPBasicBlock *HeaderVPBB = EVLPhi->getParent();
2781 VPValue *EVLIncrement = EVLPhi->getBackedgeValue();
2782 VPValue *AVL;
2783 [[maybe_unused]] bool FoundAVL =
2784 match(EVLIncrement,
2785 m_c_Add(m_ZExtOrSelf(m_EVL(m_VPValue(AVL))), m_Specific(EVLPhi)));
2786 assert(FoundAVL && "Didn't find AVL?");
2787
2788 // The AVL may be capped to a safe distance.
2789 VPValue *SafeAVL;
2790 if (match(AVL, m_Select(m_VPValue(), m_VPValue(SafeAVL), m_VPValue())))
2791 AVL = SafeAVL;
2792
2793 VPValue *AVLNext;
2794 [[maybe_unused]] bool FoundAVLNext =
2796 m_Specific(Plan.getTripCount()), m_VPValue(AVLNext)));
2797 assert(FoundAVLNext && "Didn't find AVL backedge?");
2798
2799 // Convert EVLPhi to concrete recipe.
2800 auto *ScalarR =
2801 VPBuilder(EVLPhi).createScalarPhi({EVLPhi->getStartValue(), EVLIncrement},
2802 EVLPhi->getDebugLoc(), "evl.based.iv");
2803 EVLPhi->replaceAllUsesWith(ScalarR);
2804 EVLPhi->eraseFromParent();
2805
2806 // Replace CanonicalIVInc with EVL-PHI increment.
2807 auto *CanonicalIV = cast<VPPhi>(&*HeaderVPBB->begin());
2808 VPValue *Backedge = CanonicalIV->getIncomingValue(1);
2809 assert(match(Backedge, m_c_Add(m_Specific(CanonicalIV),
2810 m_Specific(&Plan.getVFxUF()))) &&
2811 "Unexpected canonical iv");
2812 Backedge->replaceAllUsesWith(EVLIncrement);
2813
2814 // Remove unused phi and increment.
2815 VPRecipeBase *CanonicalIVIncrement = Backedge->getDefiningRecipe();
2816 CanonicalIVIncrement->eraseFromParent();
2817 CanonicalIV->eraseFromParent();
2818
2819 // Replace the use of VectorTripCount in the latch-exiting block.
2820 // Before: (branch-on-count EVLIVInc, VectorTripCount)
2821 // After: (branch-on-cond eq AVLNext, 0)
2822
2823 VPBasicBlock *LatchExiting =
2824 HeaderVPBB->getPredecessors()[1]->getEntryBasicBlock();
2825 auto *LatchExitingBr = cast<VPInstruction>(LatchExiting->getTerminator());
2826 // Skip single-iteration loop region
2827 if (match(LatchExitingBr, m_BranchOnCond(m_True())))
2828 return;
2829 assert(LatchExitingBr &&
2830 match(LatchExitingBr,
2831 m_BranchOnCount(m_VPValue(EVLIncrement),
2832 m_Specific(&Plan.getVectorTripCount()))) &&
2833 "Unexpected terminator in EVL loop");
2834
2835 Type *AVLTy = VPTypeAnalysis(Plan).inferScalarType(AVLNext);
2836 VPBuilder Builder(LatchExitingBr);
2837 VPValue *Cmp =
2838 Builder.createICmp(CmpInst::ICMP_EQ, AVLNext,
2840 Builder.createNaryOp(VPInstruction::BranchOnCond, Cmp);
2841 LatchExitingBr->eraseFromParent();
2842}
2843
2845 VPlan &Plan, PredicatedScalarEvolution &PSE,
2846 const DenseMap<Value *, const SCEV *> &StridesMap) {
2847 // Replace VPValues for known constant strides guaranteed by predicate scalar
2848 // evolution.
2849 auto CanUseVersionedStride = [&Plan](VPUser &U, unsigned) {
2850 auto *R = cast<VPRecipeBase>(&U);
2851 return R->getParent()->getParent() ||
2852 R->getParent() == Plan.getVectorLoopRegion()->getSinglePredecessor();
2853 };
2854 for (const SCEV *Stride : StridesMap.values()) {
2855 using namespace SCEVPatternMatch;
2856 auto *StrideV = cast<SCEVUnknown>(Stride)->getValue();
2857 const APInt *StrideConst;
2858 if (!match(PSE.getSCEV(StrideV), m_scev_APInt(StrideConst)))
2859 // Only handle constant strides for now.
2860 continue;
2861
2862 auto *CI =
2863 Plan.getOrAddLiveIn(ConstantInt::get(Stride->getType(), *StrideConst));
2864 if (VPValue *StrideVPV = Plan.getLiveIn(StrideV))
2865 StrideVPV->replaceUsesWithIf(CI, CanUseVersionedStride);
2866
2867 // The versioned value may not be used in the loop directly but through a
2868 // sext/zext. Add new live-ins in those cases.
2869 for (Value *U : StrideV->users()) {
2871 continue;
2872 VPValue *StrideVPV = Plan.getLiveIn(U);
2873 if (!StrideVPV)
2874 continue;
2875 unsigned BW = U->getType()->getScalarSizeInBits();
2876 APInt C =
2877 isa<SExtInst>(U) ? StrideConst->sext(BW) : StrideConst->zext(BW);
2878 VPValue *CI = Plan.getOrAddLiveIn(ConstantInt::get(U->getType(), C));
2879 StrideVPV->replaceUsesWithIf(CI, CanUseVersionedStride);
2880 }
2881 }
2882}
2883
2885 VPlan &Plan,
2886 const std::function<bool(BasicBlock *)> &BlockNeedsPredication) {
2887 // Collect recipes in the backward slice of `Root` that may generate a poison
2888 // value that is used after vectorization.
2890 auto CollectPoisonGeneratingInstrsInBackwardSlice([&](VPRecipeBase *Root) {
2892 Worklist.push_back(Root);
2893
2894 // Traverse the backward slice of Root through its use-def chain.
2895 while (!Worklist.empty()) {
2896 VPRecipeBase *CurRec = Worklist.pop_back_val();
2897
2898 if (!Visited.insert(CurRec).second)
2899 continue;
2900
2901 // Prune search if we find another recipe generating a widen memory
2902 // instruction. Widen memory instructions involved in address computation
2903 // will lead to gather/scatter instructions, which don't need to be
2904 // handled.
2906 VPHeaderPHIRecipe>(CurRec))
2907 continue;
2908
2909 // This recipe contributes to the address computation of a widen
2910 // load/store. If the underlying instruction has poison-generating flags,
2911 // drop them directly.
2912 if (auto *RecWithFlags = dyn_cast<VPRecipeWithIRFlags>(CurRec)) {
2913 VPValue *A, *B;
2914 // Dropping disjoint from an OR may yield incorrect results, as some
2915 // analysis may have converted it to an Add implicitly (e.g. SCEV used
2916 // for dependence analysis). Instead, replace it with an equivalent Add.
2917 // This is possible as all users of the disjoint OR only access lanes
2918 // where the operands are disjoint or poison otherwise.
2919 if (match(RecWithFlags, m_BinaryOr(m_VPValue(A), m_VPValue(B))) &&
2920 RecWithFlags->isDisjoint()) {
2921 VPBuilder Builder(RecWithFlags);
2922 VPInstruction *New = Builder.createOverflowingOp(
2923 Instruction::Add, {A, B}, {false, false},
2924 RecWithFlags->getDebugLoc());
2925 New->setUnderlyingValue(RecWithFlags->getUnderlyingValue());
2926 RecWithFlags->replaceAllUsesWith(New);
2927 RecWithFlags->eraseFromParent();
2928 CurRec = New;
2929 } else
2930 RecWithFlags->dropPoisonGeneratingFlags();
2931 } else {
2934 (void)Instr;
2935 assert((!Instr || !Instr->hasPoisonGeneratingFlags()) &&
2936 "found instruction with poison generating flags not covered by "
2937 "VPRecipeWithIRFlags");
2938 }
2939
2940 // Add new definitions to the worklist.
2941 for (VPValue *Operand : CurRec->operands())
2942 if (VPRecipeBase *OpDef = Operand->getDefiningRecipe())
2943 Worklist.push_back(OpDef);
2944 }
2945 });
2946
2947 // Traverse all the recipes in the VPlan and collect the poison-generating
2948 // recipes in the backward slice starting at the address of a VPWidenRecipe or
2949 // VPInterleaveRecipe.
2950 auto Iter = vp_depth_first_deep(Plan.getEntry());
2952 for (VPRecipeBase &Recipe : *VPBB) {
2953 if (auto *WidenRec = dyn_cast<VPWidenMemoryRecipe>(&Recipe)) {
2954 Instruction &UnderlyingInstr = WidenRec->getIngredient();
2955 VPRecipeBase *AddrDef = WidenRec->getAddr()->getDefiningRecipe();
2956 if (AddrDef && WidenRec->isConsecutive() &&
2957 BlockNeedsPredication(UnderlyingInstr.getParent()))
2958 CollectPoisonGeneratingInstrsInBackwardSlice(AddrDef);
2959 } else if (auto *InterleaveRec = dyn_cast<VPInterleaveRecipe>(&Recipe)) {
2960 VPRecipeBase *AddrDef = InterleaveRec->getAddr()->getDefiningRecipe();
2961 if (AddrDef) {
2962 // Check if any member of the interleave group needs predication.
2963 const InterleaveGroup<Instruction> *InterGroup =
2964 InterleaveRec->getInterleaveGroup();
2965 bool NeedPredication = false;
2966 for (int I = 0, NumMembers = InterGroup->getNumMembers();
2967 I < NumMembers; ++I) {
2968 Instruction *Member = InterGroup->getMember(I);
2969 if (Member)
2970 NeedPredication |= BlockNeedsPredication(Member->getParent());
2971 }
2972
2973 if (NeedPredication)
2974 CollectPoisonGeneratingInstrsInBackwardSlice(AddrDef);
2975 }
2976 }
2977 }
2978 }
2979}
2980
2982 VPlan &Plan,
2984 &InterleaveGroups,
2985 VPRecipeBuilder &RecipeBuilder, const bool &ScalarEpilogueAllowed) {
2986 if (InterleaveGroups.empty())
2987 return;
2988
2989 // Interleave memory: for each Interleave Group we marked earlier as relevant
2990 // for this VPlan, replace the Recipes widening its memory instructions with a
2991 // single VPInterleaveRecipe at its insertion point.
2992 VPDominatorTree VPDT;
2993 VPDT.recalculate(Plan);
2994 for (const auto *IG : InterleaveGroups) {
2995 auto *Start =
2996 cast<VPWidenMemoryRecipe>(RecipeBuilder.getRecipe(IG->getMember(0)));
2997 VPIRMetadata InterleaveMD(*Start);
2998 SmallVector<VPValue *, 4> StoredValues;
2999 if (auto *StoreR = dyn_cast<VPWidenStoreRecipe>(Start))
3000 StoredValues.push_back(StoreR->getStoredValue());
3001 for (unsigned I = 1; I < IG->getFactor(); ++I) {
3002 Instruction *MemberI = IG->getMember(I);
3003 if (!MemberI)
3004 continue;
3005 VPWidenMemoryRecipe *MemoryR =
3006 cast<VPWidenMemoryRecipe>(RecipeBuilder.getRecipe(MemberI));
3007 if (auto *StoreR = dyn_cast<VPWidenStoreRecipe>(MemoryR))
3008 StoredValues.push_back(StoreR->getStoredValue());
3009 InterleaveMD.intersect(*MemoryR);
3010 }
3011
3012 bool NeedsMaskForGaps =
3013 (IG->requiresScalarEpilogue() && !ScalarEpilogueAllowed) ||
3014 (!StoredValues.empty() && !IG->isFull());
3015
3016 Instruction *IRInsertPos = IG->getInsertPos();
3017 auto *InsertPos =
3018 cast<VPWidenMemoryRecipe>(RecipeBuilder.getRecipe(IRInsertPos));
3019
3021 if (auto *Gep = dyn_cast<GetElementPtrInst>(
3022 getLoadStorePointerOperand(IRInsertPos)->stripPointerCasts()))
3023 NW = Gep->getNoWrapFlags().withoutNoUnsignedWrap();
3024
3025 // Get or create the start address for the interleave group.
3026 VPValue *Addr = Start->getAddr();
3027 VPRecipeBase *AddrDef = Addr->getDefiningRecipe();
3028 if (AddrDef && !VPDT.properlyDominates(AddrDef, InsertPos)) {
3029 // We cannot re-use the address of member zero because it does not
3030 // dominate the insert position. Instead, use the address of the insert
3031 // position and create a PtrAdd adjusting it to the address of member
3032 // zero.
3033 // TODO: Hoist Addr's defining recipe (and any operands as needed) to
3034 // InsertPos or sink loads above zero members to join it.
3035 assert(IG->getIndex(IRInsertPos) != 0 &&
3036 "index of insert position shouldn't be zero");
3037 auto &DL = IRInsertPos->getDataLayout();
3038 APInt Offset(32,
3039 DL.getTypeAllocSize(getLoadStoreType(IRInsertPos)) *
3040 IG->getIndex(IRInsertPos),
3041 /*IsSigned=*/true);
3042 VPValue *OffsetVPV =
3043 Plan.getOrAddLiveIn(ConstantInt::get(Plan.getContext(), -Offset));
3044 VPBuilder B(InsertPos);
3045 Addr = B.createNoWrapPtrAdd(InsertPos->getAddr(), OffsetVPV, NW);
3046 }
3047 // If the group is reverse, adjust the index to refer to the last vector
3048 // lane instead of the first. We adjust the index from the first vector
3049 // lane, rather than directly getting the pointer for lane VF - 1, because
3050 // the pointer operand of the interleaved access is supposed to be uniform.
3051 if (IG->isReverse()) {
3052 auto *ReversePtr = new VPVectorEndPointerRecipe(
3053 Addr, &Plan.getVF(), getLoadStoreType(IRInsertPos),
3054 -(int64_t)IG->getFactor(), NW, InsertPos->getDebugLoc());
3055 ReversePtr->insertBefore(InsertPos);
3056 Addr = ReversePtr;
3057 }
3058 auto *VPIG = new VPInterleaveRecipe(IG, Addr, StoredValues,
3059 InsertPos->getMask(), NeedsMaskForGaps,
3060 InterleaveMD, InsertPos->getDebugLoc());
3061 VPIG->insertBefore(InsertPos);
3062
3063 unsigned J = 0;
3064 for (unsigned i = 0; i < IG->getFactor(); ++i)
3065 if (Instruction *Member = IG->getMember(i)) {
3066 VPRecipeBase *MemberR = RecipeBuilder.getRecipe(Member);
3067 if (!Member->getType()->isVoidTy()) {
3068 VPValue *OriginalV = MemberR->getVPSingleValue();
3069 OriginalV->replaceAllUsesWith(VPIG->getVPValue(J));
3070 J++;
3071 }
3072 MemberR->eraseFromParent();
3073 }
3074 }
3075}
3076
3077/// Expand a VPWidenIntOrFpInduction into executable recipes, for the initial
3078/// value, phi and backedge value. In the following example:
3079///
3080/// vector.ph:
3081/// Successor(s): vector loop
3082///
3083/// <x1> vector loop: {
3084/// vector.body:
3085/// WIDEN-INDUCTION %i = phi %start, %step, %vf
3086/// ...
3087/// EMIT branch-on-count ...
3088/// No successors
3089/// }
3090///
3091/// WIDEN-INDUCTION will get expanded to:
3092///
3093/// vector.ph:
3094/// ...
3095/// vp<%induction.start> = ...
3096/// vp<%induction.increment> = ...
3097///
3098/// Successor(s): vector loop
3099///
3100/// <x1> vector loop: {
3101/// vector.body:
3102/// ir<%i> = WIDEN-PHI vp<%induction.start>, vp<%vec.ind.next>
3103/// ...
3104/// vp<%vec.ind.next> = add ir<%i>, vp<%induction.increment>
3105/// EMIT branch-on-count ...
3106/// No successors
3107/// }
3108static void
3110 VPTypeAnalysis &TypeInfo) {
3111 VPlan *Plan = WidenIVR->getParent()->getPlan();
3112 VPValue *Start = WidenIVR->getStartValue();
3113 VPValue *Step = WidenIVR->getStepValue();
3114 VPValue *VF = WidenIVR->getVFValue();
3115 DebugLoc DL = WidenIVR->getDebugLoc();
3116
3117 // The value from the original loop to which we are mapping the new induction
3118 // variable.
3119 Type *Ty = TypeInfo.inferScalarType(WidenIVR);
3120
3121 const InductionDescriptor &ID = WidenIVR->getInductionDescriptor();
3124 // FIXME: The newly created binary instructions should contain nsw/nuw
3125 // flags, which can be found from the original scalar operations.
3126 VPIRFlags Flags;
3127 if (ID.getKind() == InductionDescriptor::IK_IntInduction) {
3128 AddOp = Instruction::Add;
3129 MulOp = Instruction::Mul;
3130 } else {
3131 AddOp = ID.getInductionOpcode();
3132 MulOp = Instruction::FMul;
3133 Flags = ID.getInductionBinOp()->getFastMathFlags();
3134 }
3135
3136 // If the phi is truncated, truncate the start and step values.
3137 VPBuilder Builder(Plan->getVectorPreheader());
3138 Type *StepTy = TypeInfo.inferScalarType(Step);
3139 if (Ty->getScalarSizeInBits() < StepTy->getScalarSizeInBits()) {
3140 assert(StepTy->isIntegerTy() && "Truncation requires an integer type");
3141 Step = Builder.createScalarCast(Instruction::Trunc, Step, Ty, DL);
3142 Start = Builder.createScalarCast(Instruction::Trunc, Start, Ty, DL);
3143 StepTy = Ty;
3144 }
3145
3146 // Construct the initial value of the vector IV in the vector loop preheader.
3147 Type *IVIntTy =
3149 VPValue *Init = Builder.createNaryOp(VPInstruction::StepVector, {}, IVIntTy);
3150 if (StepTy->isFloatingPointTy())
3151 Init = Builder.createWidenCast(Instruction::UIToFP, Init, StepTy);
3152
3153 VPValue *SplatStart = Builder.createNaryOp(VPInstruction::Broadcast, Start);
3154 VPValue *SplatStep = Builder.createNaryOp(VPInstruction::Broadcast, Step);
3155
3156 Init = Builder.createNaryOp(MulOp, {Init, SplatStep}, Flags);
3157 Init = Builder.createNaryOp(AddOp, {SplatStart, Init}, Flags,
3158 DebugLoc::getUnknown(), "induction");
3159
3160 // Create the widened phi of the vector IV.
3161 auto *WidePHI = new VPWidenPHIRecipe(WidenIVR->getPHINode(), nullptr,
3162 WidenIVR->getDebugLoc(), "vec.ind");
3163 WidePHI->addOperand(Init);
3164 WidePHI->insertBefore(WidenIVR);
3165
3166 // Create the backedge value for the vector IV.
3167 VPValue *Inc;
3168 VPValue *Prev;
3169 // If unrolled, use the increment and prev value from the operands.
3170 if (auto *SplatVF = WidenIVR->getSplatVFValue()) {
3171 Inc = SplatVF;
3172 Prev = WidenIVR->getLastUnrolledPartOperand();
3173 } else {
3174 if (VPRecipeBase *R = VF->getDefiningRecipe())
3175 Builder.setInsertPoint(R->getParent(), std::next(R->getIterator()));
3176 // Multiply the vectorization factor by the step using integer or
3177 // floating-point arithmetic as appropriate.
3178 if (StepTy->isFloatingPointTy())
3179 VF = Builder.createScalarCast(Instruction::CastOps::UIToFP, VF, StepTy,
3180 DL);
3181 else
3182 VF = Builder.createScalarZExtOrTrunc(VF, StepTy,
3183 TypeInfo.inferScalarType(VF), DL);
3184
3185 Inc = Builder.createNaryOp(MulOp, {Step, VF}, Flags);
3186 Inc = Builder.createNaryOp(VPInstruction::Broadcast, Inc);
3187 Prev = WidePHI;
3188 }
3189
3191 Builder.setInsertPoint(ExitingBB, ExitingBB->getTerminator()->getIterator());
3192 auto *Next = Builder.createNaryOp(AddOp, {Prev, Inc}, Flags,
3193 WidenIVR->getDebugLoc(), "vec.ind.next");
3194
3195 WidePHI->addOperand(Next);
3196
3197 WidenIVR->replaceAllUsesWith(WidePHI);
3198}
3199
3200/// Expand a VPWidenPointerInductionRecipe into executable recipes, for the
3201/// initial value, phi and backedge value. In the following example:
3202///
3203/// <x1> vector loop: {
3204/// vector.body:
3205/// EMIT ir<%ptr.iv> = WIDEN-POINTER-INDUCTION %start, %step, %vf
3206/// ...
3207/// EMIT branch-on-count ...
3208/// }
3209///
3210/// WIDEN-POINTER-INDUCTION will get expanded to:
3211///
3212/// <x1> vector loop: {
3213/// vector.body:
3214/// EMIT-SCALAR %pointer.phi = phi %start, %ptr.ind
3215/// EMIT %mul = mul %stepvector, %step
3216/// EMIT %vector.gep = wide-ptradd %pointer.phi, %mul
3217/// ...
3218/// EMIT %ptr.ind = ptradd %pointer.phi, %vf
3219/// EMIT branch-on-count ...
3220/// }
3222 VPTypeAnalysis &TypeInfo) {
3223 VPlan *Plan = R->getParent()->getPlan();
3224 VPValue *Start = R->getStartValue();
3225 VPValue *Step = R->getStepValue();
3226 VPValue *VF = R->getVFValue();
3227
3228 assert(R->getInductionDescriptor().getKind() ==
3230 "Not a pointer induction according to InductionDescriptor!");
3231 assert(TypeInfo.inferScalarType(R)->isPointerTy() && "Unexpected type.");
3232 assert(!R->onlyScalarsGenerated(Plan->hasScalableVF()) &&
3233 "Recipe should have been replaced");
3234
3235 VPBuilder Builder(R);
3236 DebugLoc DL = R->getDebugLoc();
3237
3238 // Build a scalar pointer phi.
3239 VPPhi *ScalarPtrPhi = Builder.createScalarPhi(Start, DL, "pointer.phi");
3240
3241 // Create actual address geps that use the pointer phi as base and a
3242 // vectorized version of the step value (<step*0, ..., step*N>) as offset.
3243 Builder.setInsertPoint(R->getParent(), R->getParent()->getFirstNonPhi());
3244 Type *StepTy = TypeInfo.inferScalarType(Step);
3245 VPValue *Offset = Builder.createNaryOp(VPInstruction::StepVector, {}, StepTy);
3246 Offset = Builder.createNaryOp(Instruction::Mul, {Offset, Step});
3247 VPValue *PtrAdd = Builder.createNaryOp(
3248 VPInstruction::WidePtrAdd, {ScalarPtrPhi, Offset}, DL, "vector.gep");
3249 R->replaceAllUsesWith(PtrAdd);
3250
3251 // Create the backedge value for the scalar pointer phi.
3253 Builder.setInsertPoint(ExitingBB, ExitingBB->getTerminator()->getIterator());
3254 VF = Builder.createScalarZExtOrTrunc(VF, StepTy, TypeInfo.inferScalarType(VF),
3255 DL);
3256 VPValue *Inc = Builder.createNaryOp(Instruction::Mul, {Step, VF});
3257
3258 VPValue *InductionGEP =
3259 Builder.createPtrAdd(ScalarPtrPhi, Inc, DL, "ptr.ind");
3260 ScalarPtrPhi->addOperand(InductionGEP);
3261}
3262
3264 // Replace loop regions with explicity CFG.
3265 SmallVector<VPRegionBlock *> LoopRegions;
3267 vp_depth_first_deep(Plan.getEntry()))) {
3268 if (!R->isReplicator())
3269 LoopRegions.push_back(R);
3270 }
3271 for (VPRegionBlock *R : LoopRegions)
3272 R->dissolveToCFGLoop();
3273}
3274
3276 VPTypeAnalysis TypeInfo(Plan);
3279 vp_depth_first_deep(Plan.getEntry()))) {
3280 for (VPRecipeBase &R : make_early_inc_range(*VPBB)) {
3281 if (auto *WidenIVR = dyn_cast<VPWidenIntOrFpInductionRecipe>(&R)) {
3282 expandVPWidenIntOrFpInduction(WidenIVR, TypeInfo);
3283 ToRemove.push_back(WidenIVR);
3284 continue;
3285 }
3286
3287 if (auto *WidenIVR = dyn_cast<VPWidenPointerInductionRecipe>(&R)) {
3288 expandVPWidenPointerInduction(WidenIVR, TypeInfo);
3289 ToRemove.push_back(WidenIVR);
3290 continue;
3291 }
3292
3293 // Expand VPBlendRecipe into VPInstruction::Select.
3294 VPBuilder Builder(&R);
3295 if (auto *Blend = dyn_cast<VPBlendRecipe>(&R)) {
3296 VPValue *Select = Blend->getIncomingValue(0);
3297 for (unsigned I = 1; I != Blend->getNumIncomingValues(); ++I)
3298 Select = Builder.createSelect(Blend->getMask(I),
3299 Blend->getIncomingValue(I), Select,
3300 R.getDebugLoc(), "predphi");
3301 Blend->replaceAllUsesWith(Select);
3302 ToRemove.push_back(Blend);
3303 }
3304
3305 if (auto *Expr = dyn_cast<VPExpressionRecipe>(&R)) {
3306 Expr->decompose();
3307 ToRemove.push_back(Expr);
3308 }
3309
3310 VPValue *VectorStep;
3311 VPValue *ScalarStep;
3313 m_VPValue(VectorStep), m_VPValue(ScalarStep))))
3314 continue;
3315
3316 // Expand WideIVStep.
3317 auto *VPI = cast<VPInstruction>(&R);
3318 Type *IVTy = TypeInfo.inferScalarType(VPI);
3319 if (TypeInfo.inferScalarType(VectorStep) != IVTy) {
3321 ? Instruction::UIToFP
3322 : Instruction::Trunc;
3323 VectorStep = Builder.createWidenCast(CastOp, VectorStep, IVTy);
3324 }
3325
3326 [[maybe_unused]] auto *ConstStep =
3327 ScalarStep->isLiveIn()
3329 : nullptr;
3330 assert(!ConstStep || ConstStep->getValue() != 1);
3331 (void)ConstStep;
3332 if (TypeInfo.inferScalarType(ScalarStep) != IVTy) {
3333 ScalarStep =
3334 Builder.createWidenCast(Instruction::Trunc, ScalarStep, IVTy);
3335 }
3336
3337 VPIRFlags Flags;
3338 if (IVTy->isFloatingPointTy())
3339 Flags = {VPI->getFastMathFlags()};
3340
3341 unsigned MulOpc =
3342 IVTy->isFloatingPointTy() ? Instruction::FMul : Instruction::Mul;
3343 VPInstruction *Mul = Builder.createNaryOp(
3344 MulOpc, {VectorStep, ScalarStep}, Flags, R.getDebugLoc());
3345 VectorStep = Mul;
3346 VPI->replaceAllUsesWith(VectorStep);
3347 ToRemove.push_back(VPI);
3348 }
3349 }
3350
3351 for (VPRecipeBase *R : ToRemove)
3352 R->eraseFromParent();
3353}
3354
3356 VPBasicBlock *EarlyExitVPBB,
3357 VPlan &Plan,
3358 VPBasicBlock *HeaderVPBB,
3359 VPBasicBlock *LatchVPBB) {
3360 VPBlockBase *MiddleVPBB = LatchVPBB->getSuccessors()[0];
3361 if (!EarlyExitVPBB->getSinglePredecessor() &&
3362 EarlyExitVPBB->getPredecessors()[1] == MiddleVPBB) {
3363 assert(EarlyExitVPBB->getNumPredecessors() == 2 &&
3364 EarlyExitVPBB->getPredecessors()[0] == EarlyExitingVPBB &&
3365 "unsupported early exit VPBB");
3366 // Early exit operand should always be last phi operand. If EarlyExitVPBB
3367 // has two predecessors and EarlyExitingVPBB is the first, swap the operands
3368 // of the phis.
3369 for (VPRecipeBase &R : EarlyExitVPBB->phis())
3370 cast<VPIRPhi>(&R)->swapOperands();
3371 }
3372
3373 VPBuilder Builder(LatchVPBB->getTerminator());
3374 VPBlockBase *TrueSucc = EarlyExitingVPBB->getSuccessors()[0];
3375 assert(
3376 match(EarlyExitingVPBB->getTerminator(), m_BranchOnCond(m_VPValue())) &&
3377 "Terminator must be be BranchOnCond");
3378 VPValue *CondOfEarlyExitingVPBB =
3379 EarlyExitingVPBB->getTerminator()->getOperand(0);
3380 auto *CondToEarlyExit = TrueSucc == EarlyExitVPBB
3381 ? CondOfEarlyExitingVPBB
3382 : Builder.createNot(CondOfEarlyExitingVPBB);
3383
3384 // Split the middle block and have it conditionally branch to the early exit
3385 // block if CondToEarlyExit.
3386 VPValue *IsEarlyExitTaken =
3387 Builder.createNaryOp(VPInstruction::AnyOf, {CondToEarlyExit});
3388 VPBasicBlock *NewMiddle = Plan.createVPBasicBlock("middle.split");
3389 VPBasicBlock *VectorEarlyExitVPBB =
3390 Plan.createVPBasicBlock("vector.early.exit");
3391 VPBlockUtils::insertOnEdge(LatchVPBB, MiddleVPBB, NewMiddle);
3392 VPBlockUtils::connectBlocks(NewMiddle, VectorEarlyExitVPBB);
3393 NewMiddle->swapSuccessors();
3394
3395 VPBlockUtils::connectBlocks(VectorEarlyExitVPBB, EarlyExitVPBB);
3396
3397 // Update the exit phis in the early exit block.
3398 VPBuilder MiddleBuilder(NewMiddle);
3399 VPBuilder EarlyExitB(VectorEarlyExitVPBB);
3400 for (VPRecipeBase &R : EarlyExitVPBB->phis()) {
3401 auto *ExitIRI = cast<VPIRPhi>(&R);
3402 // Early exit operand should always be last, i.e., 0 if EarlyExitVPBB has
3403 // a single predecessor and 1 if it has two.
3404 unsigned EarlyExitIdx = ExitIRI->getNumOperands() - 1;
3405 if (ExitIRI->getNumOperands() != 1) {
3406 // The first of two operands corresponds to the latch exit, via MiddleVPBB
3407 // predecessor. Extract its last lane.
3408 ExitIRI->extractLastLaneOfFirstOperand(MiddleBuilder);
3409 }
3410
3411 VPValue *IncomingFromEarlyExit = ExitIRI->getOperand(EarlyExitIdx);
3412 if (!IncomingFromEarlyExit->isLiveIn()) {
3413 // Update the incoming value from the early exit.
3414 VPValue *FirstActiveLane = EarlyExitB.createNaryOp(
3415 VPInstruction::FirstActiveLane, {CondToEarlyExit}, nullptr,
3416 "first.active.lane");
3417 IncomingFromEarlyExit = EarlyExitB.createNaryOp(
3418 VPInstruction::ExtractLane, {FirstActiveLane, IncomingFromEarlyExit},
3419 nullptr, "early.exit.value");
3420 ExitIRI->setOperand(EarlyExitIdx, IncomingFromEarlyExit);
3421 }
3422 }
3423 MiddleBuilder.createNaryOp(VPInstruction::BranchOnCond, {IsEarlyExitTaken});
3424
3425 // Replace the condition controlling the non-early exit from the vector loop
3426 // with one exiting if either the original condition of the vector latch is
3427 // true or the early exit has been taken.
3428 auto *LatchExitingBranch = cast<VPInstruction>(LatchVPBB->getTerminator());
3429 assert(LatchExitingBranch->getOpcode() == VPInstruction::BranchOnCount &&
3430 "Unexpected terminator");
3431 auto *IsLatchExitTaken =
3432 Builder.createICmp(CmpInst::ICMP_EQ, LatchExitingBranch->getOperand(0),
3433 LatchExitingBranch->getOperand(1));
3434 auto *AnyExitTaken = Builder.createNaryOp(
3435 Instruction::Or, {IsEarlyExitTaken, IsLatchExitTaken});
3436 Builder.createNaryOp(VPInstruction::BranchOnCond, AnyExitTaken);
3437 LatchExitingBranch->eraseFromParent();
3438}
3439
3440/// This function tries convert extended in-loop reductions to
3441/// VPExpressionRecipe and clamp the \p Range if it is beneficial and
3442/// valid. The created recipe must be decomposed to its constituent
3443/// recipes before execution.
3444static VPExpressionRecipe *
3446 VFRange &Range) {
3447 Type *RedTy = Ctx.Types.inferScalarType(Red);
3448 VPValue *VecOp = Red->getVecOp();
3449
3450 // Clamp the range if using extended-reduction is profitable.
3451 auto IsExtendedRedValidAndClampRange = [&](unsigned Opcode, bool isZExt,
3452 Type *SrcTy) -> bool {
3454 [&](ElementCount VF) {
3455 auto *SrcVecTy = cast<VectorType>(toVectorTy(SrcTy, VF));
3457 InstructionCost ExtRedCost = Ctx.TTI.getExtendedReductionCost(
3458 Opcode, isZExt, RedTy, SrcVecTy, Red->getFastMathFlags(),
3459 CostKind);
3460 InstructionCost ExtCost =
3461 cast<VPWidenCastRecipe>(VecOp)->computeCost(VF, Ctx);
3462 InstructionCost RedCost = Red->computeCost(VF, Ctx);
3463 return ExtRedCost.isValid() && ExtRedCost < ExtCost + RedCost;
3464 },
3465 Range);
3466 };
3467
3468 VPValue *A;
3469 // Match reduce(ext)).
3470 if (match(VecOp, m_ZExtOrSExt(m_VPValue(A))) &&
3471 IsExtendedRedValidAndClampRange(
3472 RecurrenceDescriptor::getOpcode(Red->getRecurrenceKind()),
3473 cast<VPWidenCastRecipe>(VecOp)->getOpcode() ==
3474 Instruction::CastOps::ZExt,
3475 Ctx.Types.inferScalarType(A)))
3476 return new VPExpressionRecipe(cast<VPWidenCastRecipe>(VecOp), Red);
3477
3478 return nullptr;
3479}
3480
3481/// This function tries convert extended in-loop reductions to
3482/// VPExpressionRecipe and clamp the \p Range if it is beneficial
3483/// and valid. The created VPExpressionRecipe must be decomposed to its
3484/// constituent recipes before execution. Patterns of the
3485/// VPExpressionRecipe:
3486/// reduce.add(mul(...)),
3487/// reduce.add(mul(ext(A), ext(B))),
3488/// reduce.add(ext(mul(ext(A), ext(B)))).
3489static VPExpressionRecipe *
3491 VPCostContext &Ctx, VFRange &Range) {
3492 unsigned Opcode = RecurrenceDescriptor::getOpcode(Red->getRecurrenceKind());
3493 if (Opcode != Instruction::Add && Opcode != Instruction::Sub)
3494 return nullptr;
3495
3496 Type *RedTy = Ctx.Types.inferScalarType(Red);
3497
3498 // Clamp the range if using multiply-accumulate-reduction is profitable.
3499 auto IsMulAccValidAndClampRange =
3500 [&](bool isZExt, VPWidenRecipe *Mul, VPWidenCastRecipe *Ext0,
3501 VPWidenCastRecipe *Ext1, VPWidenCastRecipe *OuterExt) -> bool {
3503 [&](ElementCount VF) {
3505 Type *SrcTy =
3506 Ext0 ? Ctx.Types.inferScalarType(Ext0->getOperand(0)) : RedTy;
3507 auto *SrcVecTy = cast<VectorType>(toVectorTy(SrcTy, VF));
3508 InstructionCost MulAccCost = Ctx.TTI.getMulAccReductionCost(
3509 isZExt, Opcode, RedTy, SrcVecTy, CostKind);
3510 InstructionCost MulCost = Mul->computeCost(VF, Ctx);
3511 InstructionCost RedCost = Red->computeCost(VF, Ctx);
3512 InstructionCost ExtCost = 0;
3513 if (Ext0)
3514 ExtCost += Ext0->computeCost(VF, Ctx);
3515 if (Ext1)
3516 ExtCost += Ext1->computeCost(VF, Ctx);
3517 if (OuterExt)
3518 ExtCost += OuterExt->computeCost(VF, Ctx);
3519
3520 return MulAccCost.isValid() &&
3521 MulAccCost < ExtCost + MulCost + RedCost;
3522 },
3523 Range);
3524 };
3525
3526 VPValue *VecOp = Red->getVecOp();
3527 VPValue *A, *B;
3528 // Try to match reduce.add(mul(...)).
3529 if (match(VecOp, m_Mul(m_VPValue(A), m_VPValue(B)))) {
3530 auto *RecipeA =
3531 dyn_cast_if_present<VPWidenCastRecipe>(A->getDefiningRecipe());
3532 auto *RecipeB =
3533 dyn_cast_if_present<VPWidenCastRecipe>(B->getDefiningRecipe());
3534 auto *Mul = cast<VPWidenRecipe>(VecOp->getDefiningRecipe());
3535
3536 // Match reduce.add(mul(ext, ext)).
3537 if (RecipeA && RecipeB &&
3538 (RecipeA->getOpcode() == RecipeB->getOpcode() || A == B) &&
3539 match(RecipeA, m_ZExtOrSExt(m_VPValue())) &&
3540 match(RecipeB, m_ZExtOrSExt(m_VPValue())) &&
3541 IsMulAccValidAndClampRange(RecipeA->getOpcode() ==
3542 Instruction::CastOps::ZExt,
3543 Mul, RecipeA, RecipeB, nullptr)) {
3544 return new VPExpressionRecipe(RecipeA, RecipeB, Mul, Red);
3545 }
3546 // Match reduce.add(mul).
3547 if (IsMulAccValidAndClampRange(true, Mul, nullptr, nullptr, nullptr))
3548 return new VPExpressionRecipe(Mul, Red);
3549 }
3550 // Match reduce.add(ext(mul(ext(A), ext(B)))).
3551 // All extend recipes must have same opcode or A == B
3552 // which can be transform to reduce.add(zext(mul(sext(A), sext(B)))).
3554 m_ZExtOrSExt(m_VPValue()))))) {
3555 auto *Ext = cast<VPWidenCastRecipe>(VecOp->getDefiningRecipe());
3556 auto *Mul = cast<VPWidenRecipe>(Ext->getOperand(0)->getDefiningRecipe());
3557 auto *Ext0 =
3558 cast<VPWidenCastRecipe>(Mul->getOperand(0)->getDefiningRecipe());
3559 auto *Ext1 =
3560 cast<VPWidenCastRecipe>(Mul->getOperand(1)->getDefiningRecipe());
3561 if ((Ext->getOpcode() == Ext0->getOpcode() || Ext0 == Ext1) &&
3562 Ext0->getOpcode() == Ext1->getOpcode() &&
3563 IsMulAccValidAndClampRange(Ext0->getOpcode() ==
3564 Instruction::CastOps::ZExt,
3565 Mul, Ext0, Ext1, Ext)) {
3566 auto *NewExt0 = new VPWidenCastRecipe(
3567 Ext0->getOpcode(), Ext0->getOperand(0), Ext->getResultType(), *Ext0,
3568 Ext0->getDebugLoc());
3569 NewExt0->insertBefore(Ext0);
3570
3571 VPWidenCastRecipe *NewExt1 = NewExt0;
3572 if (Ext0 != Ext1) {
3573 NewExt1 = new VPWidenCastRecipe(Ext1->getOpcode(), Ext1->getOperand(0),
3574 Ext->getResultType(), *Ext1,
3575 Ext1->getDebugLoc());
3576 NewExt1->insertBefore(Ext1);
3577 }
3578 Mul->setOperand(0, NewExt0);
3579 Mul->setOperand(1, NewExt1);
3580 Red->setOperand(1, Mul);
3581 return new VPExpressionRecipe(NewExt0, NewExt1, Mul, Red);
3582 }
3583 }
3584 return nullptr;
3585}
3586
3587/// This function tries to create abstract recipes from the reduction recipe for
3588/// following optimizations and cost estimation.
3590 VPCostContext &Ctx,
3591 VFRange &Range) {
3592 VPExpressionRecipe *AbstractR = nullptr;
3593 auto IP = std::next(Red->getIterator());
3594 auto *VPBB = Red->getParent();
3595 if (auto *MulAcc = tryToMatchAndCreateMulAccumulateReduction(Red, Ctx, Range))
3596 AbstractR = MulAcc;
3597 else if (auto *ExtRed = tryToMatchAndCreateExtendedReduction(Red, Ctx, Range))
3598 AbstractR = ExtRed;
3599 // Cannot create abstract inloop reduction recipes.
3600 if (!AbstractR)
3601 return;
3602
3603 AbstractR->insertBefore(*VPBB, IP);
3604 Red->replaceAllUsesWith(AbstractR);
3605}
3606
3617
3619 if (Plan.hasScalarVFOnly())
3620 return;
3621
3622#ifndef NDEBUG
3623 VPDominatorTree VPDT;
3624 VPDT.recalculate(Plan);
3625#endif
3626
3627 SmallVector<VPValue *> VPValues;
3630 append_range(VPValues, Plan.getLiveIns());
3631 for (VPRecipeBase &R : *Plan.getEntry())
3632 append_range(VPValues, R.definedValues());
3633
3634 auto *VectorPreheader = Plan.getVectorPreheader();
3635 for (VPValue *VPV : VPValues) {
3637 (VPV->isLiveIn() && VPV->getLiveInIRValue() &&
3638 isa<Constant>(VPV->getLiveInIRValue())))
3639 continue;
3640
3641 // Add explicit broadcast at the insert point that dominates all users.
3642 VPBasicBlock *HoistBlock = VectorPreheader;
3643 VPBasicBlock::iterator HoistPoint = VectorPreheader->end();
3644 for (VPUser *User : VPV->users()) {
3645 if (User->usesScalars(VPV))
3646 continue;
3647 if (cast<VPRecipeBase>(User)->getParent() == VectorPreheader)
3648 HoistPoint = HoistBlock->begin();
3649 else
3650 assert(VPDT.dominates(VectorPreheader,
3651 cast<VPRecipeBase>(User)->getParent()) &&
3652 "All users must be in the vector preheader or dominated by it");
3653 }
3654
3655 VPBuilder Builder(cast<VPBasicBlock>(HoistBlock), HoistPoint);
3656 auto *Broadcast = Builder.createNaryOp(VPInstruction::Broadcast, {VPV});
3657 VPV->replaceUsesWithIf(Broadcast,
3658 [VPV, Broadcast](VPUser &U, unsigned Idx) {
3659 return Broadcast != &U && !U.usesScalars(VPV);
3660 });
3661 }
3662}
3663
3665 VPlan &Plan, ElementCount BestVF, unsigned BestUF,
3667 assert(Plan.hasVF(BestVF) && "BestVF is not available in Plan");
3668 assert(Plan.hasUF(BestUF) && "BestUF is not available in Plan");
3669
3670 VPValue *TC = Plan.getTripCount();
3671 // Skip cases for which the trip count may be non-trivial to materialize.
3672 // I.e., when a scalar tail is absent - due to tail folding, or when a scalar
3673 // tail is required.
3674 if (!Plan.hasScalarTail() ||
3676 Plan.getScalarPreheader() ||
3677 !TC->isLiveIn())
3678 return;
3679
3680 // Materialize vector trip counts for constants early if it can simply
3681 // be computed as (Original TC / VF * UF) * VF * UF.
3682 // TODO: Compute vector trip counts for loops requiring a scalar epilogue and
3683 // tail-folded loops.
3684 ScalarEvolution &SE = *PSE.getSE();
3685 auto *TCScev = SE.getSCEV(TC->getLiveInIRValue());
3686 if (!isa<SCEVConstant>(TCScev))
3687 return;
3688 const SCEV *VFxUF = SE.getElementCount(TCScev->getType(), BestVF * BestUF);
3689 auto VecTCScev = SE.getMulExpr(SE.getUDivExpr(TCScev, VFxUF), VFxUF);
3690 if (auto *ConstVecTC = dyn_cast<SCEVConstant>(VecTCScev))
3691 Plan.getVectorTripCount().setUnderlyingValue(ConstVecTC->getValue());
3692}
3693
3695 VPBasicBlock *VectorPH) {
3697 if (BTC->getNumUsers() == 0)
3698 return;
3699
3700 VPBuilder Builder(VectorPH, VectorPH->begin());
3701 auto *TCTy = VPTypeAnalysis(Plan).inferScalarType(Plan.getTripCount());
3702 auto *TCMO = Builder.createNaryOp(
3703 Instruction::Sub,
3704 {Plan.getTripCount(), Plan.getOrAddLiveIn(ConstantInt::get(TCTy, 1))},
3705 DebugLoc::getCompilerGenerated(), "trip.count.minus.1");
3706 BTC->replaceAllUsesWith(TCMO);
3707}
3708
3710 if (Plan.hasScalarVFOnly())
3711 return;
3712
3713 VPTypeAnalysis TypeInfo(Plan);
3714 VPRegionBlock *LoopRegion = Plan.getVectorLoopRegion();
3715 auto VPBBsOutsideLoopRegion = VPBlockUtils::blocksOnly<VPBasicBlock>(
3717 auto VPBBsInsideLoopRegion = VPBlockUtils::blocksOnly<VPBasicBlock>(
3718 vp_depth_first_shallow(LoopRegion->getEntry()));
3719 // Materialize Build(Struct)Vector for all replicating VPReplicateRecipes and
3720 // VPInstructions, excluding ones in replicate regions. Those are not
3721 // materialized explicitly yet. Those vector users are still handled in
3722 // VPReplicateRegion::execute(), via shouldPack().
3723 // TODO: materialize build vectors for replicating recipes in replicating
3724 // regions.
3725 for (VPBasicBlock *VPBB :
3726 concat<VPBasicBlock *>(VPBBsOutsideLoopRegion, VPBBsInsideLoopRegion)) {
3727 for (VPRecipeBase &R : make_early_inc_range(*VPBB)) {
3729 continue;
3730 auto *DefR = cast<VPRecipeWithIRFlags>(&R);
3731 auto UsesVectorOrInsideReplicateRegion = [DefR, LoopRegion](VPUser *U) {
3732 VPRegionBlock *ParentRegion =
3734 return !U->usesScalars(DefR) || ParentRegion != LoopRegion;
3735 };
3736 if ((isa<VPReplicateRecipe>(DefR) &&
3737 cast<VPReplicateRecipe>(DefR)->isSingleScalar()) ||
3738 (isa<VPInstruction>(DefR) &&
3740 !cast<VPInstruction>(DefR)->doesGeneratePerAllLanes())) ||
3741 none_of(DefR->users(), UsesVectorOrInsideReplicateRegion))
3742 continue;
3743
3744 Type *ScalarTy = TypeInfo.inferScalarType(DefR);
3745 unsigned Opcode = ScalarTy->isStructTy()
3748 auto *BuildVector = new VPInstruction(Opcode, {DefR});
3749 BuildVector->insertAfter(DefR);
3750
3751 DefR->replaceUsesWithIf(
3752 BuildVector, [BuildVector, &UsesVectorOrInsideReplicateRegion](
3753 VPUser &U, unsigned) {
3754 return &U != BuildVector && UsesVectorOrInsideReplicateRegion(&U);
3755 });
3756 }
3757 }
3758}
3759
3761 VPBasicBlock *VectorPHVPBB,
3762 bool TailByMasking,
3763 bool RequiresScalarEpilogue) {
3764 VPValue &VectorTC = Plan.getVectorTripCount();
3765 assert(VectorTC.isLiveIn() && "vector-trip-count must be a live-in");
3766 // There's nothing to do if there are no users of the vector trip count or its
3767 // IR value has already been set.
3768 if (VectorTC.getNumUsers() == 0 || VectorTC.getLiveInIRValue())
3769 return;
3770
3771 VPValue *TC = Plan.getTripCount();
3772 Type *TCTy = VPTypeAnalysis(Plan).inferScalarType(TC);
3773 VPBuilder Builder(VectorPHVPBB, VectorPHVPBB->begin());
3774 VPValue *Step = &Plan.getVFxUF();
3775
3776 // If the tail is to be folded by masking, round the number of iterations N
3777 // up to a multiple of Step instead of rounding down. This is done by first
3778 // adding Step-1 and then rounding down. Note that it's ok if this addition
3779 // overflows: the vector induction variable will eventually wrap to zero given
3780 // that it starts at zero and its Step is a power of two; the loop will then
3781 // exit, with the last early-exit vector comparison also producing all-true.
3782 // For scalable vectors the VF is not guaranteed to be a power of 2, but this
3783 // is accounted for in emitIterationCountCheck that adds an overflow check.
3784 if (TailByMasking) {
3785 TC = Builder.createNaryOp(
3786 Instruction::Add,
3787 {TC, Builder.createNaryOp(
3788 Instruction::Sub,
3789 {Step, Plan.getOrAddLiveIn(ConstantInt::get(TCTy, 1))})},
3790 DebugLoc::getCompilerGenerated(), "n.rnd.up");
3791 }
3792
3793 // Now we need to generate the expression for the part of the loop that the
3794 // vectorized body will execute. This is equal to N - (N % Step) if scalar
3795 // iterations are not required for correctness, or N - Step, otherwise. Step
3796 // is equal to the vectorization factor (number of SIMD elements) times the
3797 // unroll factor (number of SIMD instructions).
3798 VPValue *R =
3799 Builder.createNaryOp(Instruction::URem, {TC, Step},
3800 DebugLoc::getCompilerGenerated(), "n.mod.vf");
3801
3802 // There are cases where we *must* run at least one iteration in the remainder
3803 // loop. See the cost model for when this can happen. If the step evenly
3804 // divides the trip count, we set the remainder to be equal to the step. If
3805 // the step does not evenly divide the trip count, no adjustment is necessary
3806 // since there will already be scalar iterations. Note that the minimum
3807 // iterations check ensures that N >= Step.
3808 if (RequiresScalarEpilogue) {
3809 assert(!TailByMasking &&
3810 "requiring scalar epilogue is not supported with fail folding");
3811 VPValue *IsZero = Builder.createICmp(
3812 CmpInst::ICMP_EQ, R, Plan.getOrAddLiveIn(ConstantInt::get(TCTy, 0)));
3813 R = Builder.createSelect(IsZero, Step, R);
3814 }
3815
3816 VPValue *Res = Builder.createNaryOp(
3817 Instruction::Sub, {TC, R}, DebugLoc::getCompilerGenerated(), "n.vec");
3818 VectorTC.replaceAllUsesWith(Res);
3819}
3820
3822 ElementCount VFEC) {
3823 VPBuilder Builder(VectorPH, VectorPH->begin());
3824 Type *TCTy = VPTypeAnalysis(Plan).inferScalarType(Plan.getTripCount());
3825 VPValue &VF = Plan.getVF();
3826 VPValue &VFxUF = Plan.getVFxUF();
3827 // Note that after the transform, Plan.getVF and Plan.getVFxUF should not be
3828 // used.
3829 // TODO: Assert that they aren't used.
3830
3831 // If there are no users of the runtime VF, compute VFxUF by constant folding
3832 // the multiplication of VF and UF.
3833 if (VF.getNumUsers() == 0) {
3834 VPValue *RuntimeVFxUF =
3835 Builder.createElementCount(TCTy, VFEC * Plan.getUF());
3836 VFxUF.replaceAllUsesWith(RuntimeVFxUF);
3837 return;
3838 }
3839
3840 // For users of the runtime VF, compute it as VF * vscale, and VFxUF as (VF *
3841 // vscale) * UF.
3842 VPValue *RuntimeVF = Builder.createElementCount(TCTy, VFEC);
3844 VPValue *BC = Builder.createNaryOp(VPInstruction::Broadcast, RuntimeVF);
3846 BC, [&VF](VPUser &U, unsigned) { return !U.usesScalars(&VF); });
3847 }
3848 VF.replaceAllUsesWith(RuntimeVF);
3849
3850 VPValue *UF = Plan.getOrAddLiveIn(ConstantInt::get(TCTy, Plan.getUF()));
3851 VPValue *MulByUF = Builder.createNaryOp(Instruction::Mul, {RuntimeVF, UF});
3852 VFxUF.replaceAllUsesWith(MulByUF);
3853}
3854
3857 const DataLayout &DL = SE.getDataLayout();
3858 SCEVExpander Expander(SE, DL, "induction", /*PreserveLCSSA=*/true);
3859
3860 auto *Entry = cast<VPIRBasicBlock>(Plan.getEntry());
3861 BasicBlock *EntryBB = Entry->getIRBasicBlock();
3862 DenseMap<const SCEV *, Value *> ExpandedSCEVs;
3863 for (VPRecipeBase &R : make_early_inc_range(*Entry)) {
3865 continue;
3866 auto *ExpSCEV = dyn_cast<VPExpandSCEVRecipe>(&R);
3867 if (!ExpSCEV)
3868 break;
3869 const SCEV *Expr = ExpSCEV->getSCEV();
3870 Value *Res =
3871 Expander.expandCodeFor(Expr, Expr->getType(), EntryBB->getTerminator());
3872 ExpandedSCEVs[ExpSCEV->getSCEV()] = Res;
3873 VPValue *Exp = Plan.getOrAddLiveIn(Res);
3874 ExpSCEV->replaceAllUsesWith(Exp);
3875 if (Plan.getTripCount() == ExpSCEV)
3876 Plan.resetTripCount(Exp);
3877 ExpSCEV->eraseFromParent();
3878 }
3880 "VPExpandSCEVRecipes must be at the beginning of the entry block, "
3881 "after any VPIRInstructions");
3882 // Add IR instructions in the entry basic block but not in the VPIRBasicBlock
3883 // to the VPIRBasicBlock.
3884 auto EI = Entry->begin();
3885 for (Instruction &I : drop_end(*EntryBB)) {
3886 if (EI != Entry->end() && isa<VPIRInstruction>(*EI) &&
3887 &cast<VPIRInstruction>(&*EI)->getInstruction() == &I) {
3888 EI++;
3889 continue;
3890 }
3892 }
3893
3894 return ExpandedSCEVs;
3895}
3896
3897/// Returns true if \p V is VPWidenLoadRecipe or VPInterleaveRecipe that can be
3898/// converted to a narrower recipe. \p V is used by a wide recipe that feeds a
3899/// store interleave group at index \p Idx, \p WideMember0 is the recipe feeding
3900/// the same interleave group at index 0. A VPWidenLoadRecipe can be narrowed to
3901/// an index-independent load if it feeds all wide ops at all indices (\p OpV
3902/// must be the operand at index \p OpIdx for both the recipe at lane 0, \p
3903/// WideMember0). A VPInterleaveRecipe can be narrowed to a wide load, if \p V
3904/// is defined at \p Idx of a load interleave group.
3905static bool canNarrowLoad(VPWidenRecipe *WideMember0, unsigned OpIdx,
3906 VPValue *OpV, unsigned Idx) {
3907 auto *DefR = OpV->getDefiningRecipe();
3908 if (!DefR)
3909 return WideMember0->getOperand(OpIdx) == OpV;
3910 if (auto *W = dyn_cast<VPWidenLoadRecipe>(DefR))
3911 return !W->getMask() && WideMember0->getOperand(OpIdx) == OpV;
3912
3913 if (auto *IR = dyn_cast<VPInterleaveRecipe>(DefR))
3914 return IR->getInterleaveGroup()->isFull() && IR->getVPValue(Idx) == OpV;
3915 return false;
3916}
3917
3918/// Returns true if \p IR is a full interleave group with factor and number of
3919/// members both equal to \p VF. The interleave group must also access the full
3920/// vector width \p VectorRegWidth.
3922 unsigned VF, VPTypeAnalysis &TypeInfo,
3923 unsigned VectorRegWidth) {
3924 if (!InterleaveR)
3925 return false;
3926
3927 Type *GroupElementTy = nullptr;
3928 if (InterleaveR->getStoredValues().empty()) {
3929 GroupElementTy = TypeInfo.inferScalarType(InterleaveR->getVPValue(0));
3930 if (!all_of(InterleaveR->definedValues(),
3931 [&TypeInfo, GroupElementTy](VPValue *Op) {
3932 return TypeInfo.inferScalarType(Op) == GroupElementTy;
3933 }))
3934 return false;
3935 } else {
3936 GroupElementTy =
3937 TypeInfo.inferScalarType(InterleaveR->getStoredValues()[0]);
3938 if (!all_of(InterleaveR->getStoredValues(),
3939 [&TypeInfo, GroupElementTy](VPValue *Op) {
3940 return TypeInfo.inferScalarType(Op) == GroupElementTy;
3941 }))
3942 return false;
3943 }
3944
3945 unsigned GroupSize = GroupElementTy->getScalarSizeInBits() * VF;
3946 auto IG = InterleaveR->getInterleaveGroup();
3947 return IG->getFactor() == VF && IG->getNumMembers() == VF &&
3948 GroupSize == VectorRegWidth;
3949}
3950
3951/// Returns true if \p VPValue is a narrow VPValue.
3952static bool isAlreadyNarrow(VPValue *VPV) {
3953 if (VPV->isLiveIn())
3954 return true;
3955 auto *RepR = dyn_cast<VPReplicateRecipe>(VPV);
3956 return RepR && RepR->isSingleScalar();
3957}
3958
3960 unsigned VectorRegWidth) {
3961 VPRegionBlock *VectorLoop = Plan.getVectorLoopRegion();
3962 if (!VectorLoop)
3963 return;
3964
3965 VPTypeAnalysis TypeInfo(Plan);
3966
3967 unsigned VFMinVal = VF.getKnownMinValue();
3969 for (auto &R : *VectorLoop->getEntryBasicBlock()) {
3972 continue;
3973
3976 continue;
3977
3978 // Bail out on recipes not supported at the moment:
3979 // * phi recipes other than the canonical induction
3980 // * recipes writing to memory except interleave groups
3981 // Only support plans with a canonical induction phi.
3982 if (R.isPhi())
3983 return;
3984
3985 auto *InterleaveR = dyn_cast<VPInterleaveRecipe>(&R);
3986 if (R.mayWriteToMemory() && !InterleaveR)
3987 return;
3988
3989 // Do not narrow interleave groups if there are VectorPointer recipes and
3990 // the plan was unrolled. The recipe implicitly uses VF from
3991 // VPTransformState.
3992 // TODO: Remove restriction once the VF for the VectorPointer offset is
3993 // modeled explicitly as operand.
3994 if (isa<VPVectorPointerRecipe>(&R) && Plan.getUF() > 1)
3995 return;
3996
3997 // All other ops are allowed, but we reject uses that cannot be converted
3998 // when checking all allowed consumers (store interleave groups) below.
3999 if (!InterleaveR)
4000 continue;
4001
4002 // Bail out on non-consecutive interleave groups.
4003 if (!isConsecutiveInterleaveGroup(InterleaveR, VFMinVal, TypeInfo,
4004 VectorRegWidth))
4005 return;
4006
4007 // Skip read interleave groups.
4008 if (InterleaveR->getStoredValues().empty())
4009 continue;
4010
4011 // Narrow interleave groups, if all operands are already matching narrow
4012 // ops.
4013 auto *Member0 = InterleaveR->getStoredValues()[0];
4014 if (isAlreadyNarrow(Member0) &&
4015 all_of(InterleaveR->getStoredValues(),
4016 [Member0](VPValue *VPV) { return Member0 == VPV; })) {
4017 StoreGroups.push_back(InterleaveR);
4018 continue;
4019 }
4020
4021 // For now, we only support full interleave groups storing load interleave
4022 // groups.
4023 if (all_of(enumerate(InterleaveR->getStoredValues()), [](auto Op) {
4024 VPRecipeBase *DefR = Op.value()->getDefiningRecipe();
4025 if (!DefR)
4026 return false;
4027 auto *IR = dyn_cast<VPInterleaveRecipe>(DefR);
4028 return IR && IR->getInterleaveGroup()->isFull() &&
4029 IR->getVPValue(Op.index()) == Op.value();
4030 })) {
4031 StoreGroups.push_back(InterleaveR);
4032 continue;
4033 }
4034
4035 // Check if all values feeding InterleaveR are matching wide recipes, which
4036 // operands that can be narrowed.
4037 auto *WideMember0 = dyn_cast_or_null<VPWidenRecipe>(
4038 InterleaveR->getStoredValues()[0]->getDefiningRecipe());
4039 if (!WideMember0)
4040 return;
4041 for (const auto &[I, V] : enumerate(InterleaveR->getStoredValues())) {
4042 auto *R = dyn_cast_or_null<VPWidenRecipe>(V->getDefiningRecipe());
4043 if (!R || R->getOpcode() != WideMember0->getOpcode() ||
4044 R->getNumOperands() > 2)
4045 return;
4046 if (any_of(enumerate(R->operands()),
4047 [WideMember0, Idx = I](const auto &P) {
4048 const auto &[OpIdx, OpV] = P;
4049 return !canNarrowLoad(WideMember0, OpIdx, OpV, Idx);
4050 }))
4051 return;
4052 }
4053 StoreGroups.push_back(InterleaveR);
4054 }
4055
4056 if (StoreGroups.empty())
4057 return;
4058
4059 // Convert InterleaveGroup \p R to a single VPWidenLoadRecipe.
4060 SmallPtrSet<VPValue *, 4> NarrowedOps;
4061 auto NarrowOp = [&NarrowedOps](VPValue *V) -> VPValue * {
4062 auto *R = V->getDefiningRecipe();
4063 if (!R || NarrowedOps.contains(V))
4064 return V;
4065 if (auto *LoadGroup = dyn_cast<VPInterleaveRecipe>(R)) {
4066 // Narrow interleave group to wide load, as transformed VPlan will only
4067 // process one original iteration.
4068 auto *L = new VPWidenLoadRecipe(
4069 *cast<LoadInst>(LoadGroup->getInterleaveGroup()->getInsertPos()),
4070 LoadGroup->getAddr(), LoadGroup->getMask(), /*Consecutive=*/true,
4071 /*Reverse=*/false, {}, LoadGroup->getDebugLoc());
4072 L->insertBefore(LoadGroup);
4073 NarrowedOps.insert(L);
4074 return L;
4075 }
4076
4077 if (auto *RepR = dyn_cast<VPReplicateRecipe>(R)) {
4078 assert(RepR->isSingleScalar() &&
4079 isa<LoadInst>(RepR->getUnderlyingInstr()) &&
4080 "must be a single scalar load");
4081 NarrowedOps.insert(RepR);
4082 return RepR;
4083 }
4084 auto *WideLoad = cast<VPWidenLoadRecipe>(R);
4085
4086 VPValue *PtrOp = WideLoad->getAddr();
4087 if (auto *VecPtr = dyn_cast<VPVectorPointerRecipe>(PtrOp))
4088 PtrOp = VecPtr->getOperand(0);
4089 // Narrow wide load to uniform scalar load, as transformed VPlan will only
4090 // process one original iteration.
4091 auto *N = new VPReplicateRecipe(&WideLoad->getIngredient(), {PtrOp},
4092 /*IsUniform*/ true,
4093 /*Mask*/ nullptr, *WideLoad);
4094 N->insertBefore(WideLoad);
4095 NarrowedOps.insert(N);
4096 return N;
4097 };
4098
4099 // Narrow operation tree rooted at store groups.
4100 for (auto *StoreGroup : StoreGroups) {
4101 VPValue *Res = nullptr;
4102 VPValue *Member0 = StoreGroup->getStoredValues()[0];
4103 if (isAlreadyNarrow(Member0)) {
4104 Res = Member0;
4105 } else if (auto *WideMember0 =
4107 for (unsigned Idx = 0, E = WideMember0->getNumOperands(); Idx != E; ++Idx)
4108 WideMember0->setOperand(Idx, NarrowOp(WideMember0->getOperand(Idx)));
4109 Res = WideMember0;
4110 } else {
4111 Res = NarrowOp(Member0);
4112 }
4113
4114 auto *S = new VPWidenStoreRecipe(
4115 *cast<StoreInst>(StoreGroup->getInterleaveGroup()->getInsertPos()),
4116 StoreGroup->getAddr(), Res, nullptr, /*Consecutive=*/true,
4117 /*Reverse=*/false, {}, StoreGroup->getDebugLoc());
4118 S->insertBefore(StoreGroup);
4119 StoreGroup->eraseFromParent();
4120 }
4121
4122 // Adjust induction to reflect that the transformed plan only processes one
4123 // original iteration.
4124 auto *CanIV = Plan.getCanonicalIV();
4125 auto *Inc = cast<VPInstruction>(CanIV->getBackedgeValue());
4126 VPBuilder PHBuilder(Plan.getVectorPreheader());
4127
4128 VPValue *UF = Plan.getOrAddLiveIn(
4129 ConstantInt::get(CanIV->getScalarType(), 1 * Plan.getUF()));
4130 if (VF.isScalable()) {
4131 VPValue *VScale = PHBuilder.createElementCount(
4132 CanIV->getScalarType(), ElementCount::getScalable(1));
4133 VPValue *VScaleUF = PHBuilder.createNaryOp(Instruction::Mul, {VScale, UF});
4134 Inc->setOperand(1, VScaleUF);
4135 Plan.getVF().replaceAllUsesWith(VScale);
4136 } else {
4137 Inc->setOperand(1, UF);
4139 Plan.getOrAddLiveIn(ConstantInt::get(CanIV->getScalarType(), 1)));
4140 }
4141 removeDeadRecipes(Plan);
4142}
4143
4144/// Add branch weight metadata, if the \p Plan's middle block is terminated by a
4145/// BranchOnCond recipe.
4147 VPlan &Plan, ElementCount VF, std::optional<unsigned> VScaleForTuning) {
4148 VPBasicBlock *MiddleVPBB = Plan.getMiddleBlock();
4149 auto *MiddleTerm =
4151 // Only add branch metadata if there is a (conditional) terminator.
4152 if (!MiddleTerm)
4153 return;
4154
4155 assert(MiddleTerm->getOpcode() == VPInstruction::BranchOnCond &&
4156 "must have a BranchOnCond");
4157 // Assume that `TripCount % VectorStep ` is equally distributed.
4158 unsigned VectorStep = Plan.getUF() * VF.getKnownMinValue();
4159 if (VF.isScalable() && VScaleForTuning.has_value())
4160 VectorStep *= *VScaleForTuning;
4161 assert(VectorStep > 0 && "trip count should not be zero");
4162 MDBuilder MDB(Plan.getContext());
4163 MDNode *BranchWeights =
4164 MDB.createBranchWeights({1, VectorStep - 1}, /*IsExpected=*/false);
4165 MiddleTerm->addMetadata(LLVMContext::MD_prof, BranchWeights);
4166}
assert(UImm &&(UImm !=~static_cast< T >(0)) &&"Invalid immediate!")
AMDGPU Register Bank Select
This file implements a class to represent arbitrary precision integral constant values and operations...
ReachingDefAnalysis InstSet & ToRemove
MachineBasicBlock MachineBasicBlock::iterator DebugLoc DL
static bool isEqual(const Function &Caller, const Function &Callee)
static const Function * getParent(const Value *V)
static GCRegistry::Add< ErlangGC > A("erlang", "erlang-compatible garbage collector")
static GCRegistry::Add< OcamlGC > B("ocaml", "ocaml 3.10-compatible GC")
static cl::opt< OutputCostKind > CostKind("cost-kind", cl::desc("Target cost kind"), cl::init(OutputCostKind::RecipThroughput), cl::values(clEnumValN(OutputCostKind::RecipThroughput, "throughput", "Reciprocal throughput"), clEnumValN(OutputCostKind::Latency, "latency", "Instruction latency"), clEnumValN(OutputCostKind::CodeSize, "code-size", "Code size"), clEnumValN(OutputCostKind::SizeAndLatency, "size-latency", "Code size and latency"), clEnumValN(OutputCostKind::All, "all", "Print all cost kinds")))
static bool isSentinel(const DWARFDebugNames::AttributeEncoding &AE)
@ Default
Hexagon Common GEP
iv Induction Variable Users
Definition IVUsers.cpp:48
const AbstractManglingParser< Derived, Alloc >::OperatorInfo AbstractManglingParser< Derived, Alloc >::Ops[]
licm
Definition LICM.cpp:381
Legalize the Machine IR a function s Machine IR
Definition Legalizer.cpp:80
static bool mergeBlocksIntoPredecessors(Loop &L, DominatorTree &DT, LoopInfo &LI, MemorySSAUpdater *MSSAU, ScalarEvolution &SE)
#define I(x, y, z)
Definition MD5.cpp:58
mir Rename Register Operands
MachineInstr unsigned OpIdx
ConstantRange Range(APInt(BitWidth, Low), APInt(BitWidth, High))
uint64_t IntrinsicInst * II
#define P(N)
This file builds on the ADT/GraphTraits.h file to build a generic graph post order iterator.
const SmallVectorImpl< MachineOperand > & Cond
This file contains some templates that are useful if you are working with the STL at all.
This file implements a set that has insertion order iteration characteristics.
static TableGen::Emitter::Opt Y("gen-skeleton-entry", EmitSkeleton, "Generate example skeleton entry")
static TableGen::Emitter::OptClass< SkeletonEmitter > X("gen-skeleton-class", "Generate example skeleton class")
static SymbolRef::Type getType(const Symbol *Sym)
Definition TapiFile.cpp:39
This file implements the TypeSwitch template, which mimics a switch() statement whose cases are type ...
This file implements dominator tree analysis for a single level of a VPlan's H-CFG.
This file contains the declarations of different VPlan-related auxiliary helpers.
static VPValue * optimizeLatchExitInductionUser(VPlan &Plan, VPTypeAnalysis &TypeInfo, VPBlockBase *PredVPBB, VPValue *Op, DenseMap< VPValue *, VPValue * > &EndValues, ScalarEvolution &SE)
Attempts to optimize the induction variable exit values for users in the exit block coming from the l...
static void removeCommonBlendMask(VPBlendRecipe *Blend)
Try to see if all of Blend's masks share a common value logically and'ed and remove it from the masks...
static void tryToCreateAbstractReductionRecipe(VPReductionRecipe *Red, VPCostContext &Ctx, VFRange &Range)
This function tries to create abstract recipes from the reduction recipe for following optimizations ...
static bool sinkScalarOperands(VPlan &Plan)
static bool simplifyBranchConditionForVFAndUF(VPlan &Plan, ElementCount BestVF, unsigned BestUF, PredicatedScalarEvolution &PSE)
Try to simplify the branch condition of Plan.
static Value * tryToFoldLiveIns(const VPRecipeBase &R, unsigned Opcode, ArrayRef< VPValue * > Operands, const DataLayout &DL, VPTypeAnalysis &TypeInfo)
Try to fold R using InstSimplifyFolder.
static void removeRedundantInductionCasts(VPlan &Plan)
Remove redundant casts of inductions.
static bool tryToReplaceALMWithWideALM(VPlan &Plan, ElementCount VF, unsigned UF)
Try to replace multiple active lane masks used for control flow with a single, wide active lane mask ...
static VPExpressionRecipe * tryToMatchAndCreateExtendedReduction(VPReductionRecipe *Red, VPCostContext &Ctx, VFRange &Range)
This function tries convert extended in-loop reductions to VPExpressionRecipe and clamp the Range if ...
static VPScalarIVStepsRecipe * createScalarIVSteps(VPlan &Plan, InductionDescriptor::InductionKind Kind, Instruction::BinaryOps InductionOpcode, FPMathOperator *FPBinOp, Instruction *TruncI, VPValue *StartV, VPValue *Step, DebugLoc DL, VPBuilder &Builder)
static bool sinkRecurrenceUsersAfterPrevious(VPFirstOrderRecurrencePHIRecipe *FOR, VPRecipeBase *Previous, VPDominatorTree &VPDT)
Sink users of FOR after the recipe defining the previous value Previous of the recurrence.
static bool mergeReplicateRegionsIntoSuccessors(VPlan &Plan)
static VPActiveLaneMaskPHIRecipe * addVPLaneMaskPhiAndUpdateExitBranch(VPlan &Plan, bool DataAndControlFlowWithoutRuntimeCheck)
static void expandVPWidenPointerInduction(VPWidenPointerInductionRecipe *R, VPTypeAnalysis &TypeInfo)
Expand a VPWidenPointerInductionRecipe into executable recipes, for the initial value,...
static void transformRecipestoEVLRecipes(VPlan &Plan, VPValue &EVL)
Replace recipes with their EVL variants.
static bool isDeadRecipe(VPRecipeBase &R)
Returns true if R is dead and can be removed.
static void legalizeAndOptimizeInductions(VPlan &Plan)
Legalize VPWidenPointerInductionRecipe, by replacing it with a PtrAdd (IndStart, ScalarIVSteps (0,...
static void addReplicateRegions(VPlan &Plan)
static void simplifyRecipe(VPRecipeBase &R, VPTypeAnalysis &TypeInfo)
Try to simplify recipe R.
static void removeRedundantExpandSCEVRecipes(VPlan &Plan)
Remove redundant EpxandSCEVRecipes in Plan's entry block by replacing them with already existing reci...
static bool isConditionTrueViaVFAndUF(VPValue *Cond, VPlan &Plan, ElementCount BestVF, unsigned BestUF, ScalarEvolution &SE)
Return true if Cond is known to be true for given BestVF and BestUF.
static bool isConsecutiveInterleaveGroup(VPInterleaveRecipe *InterleaveR, unsigned VF, VPTypeAnalysis &TypeInfo, unsigned VectorRegWidth)
Returns true if IR is a full interleave group with factor and number of members both equal to VF.
static bool hoistPreviousBeforeFORUsers(VPFirstOrderRecurrencePHIRecipe *FOR, VPRecipeBase *Previous, VPDominatorTree &VPDT)
Try to hoist Previous and its operands before all users of FOR.
static SmallVector< VPUser * > collectUsersRecursively(VPValue *V)
static void recursivelyDeleteDeadRecipes(VPValue *V)
static VPValue * optimizeEarlyExitInductionUser(VPlan &Plan, VPTypeAnalysis &TypeInfo, VPBlockBase *PredVPBB, VPValue *Op, ScalarEvolution &SE)
Attempts to optimize the induction variable exit values for users in the early exit block.
cl::opt< bool > EnableWideActiveLaneMask("enable-wide-lane-mask", cl::init(false), cl::Hidden, cl::desc("Enable use of wide get active lane mask instructions"))
static VPWidenInductionRecipe * getOptimizableIVOf(VPValue *VPV, ScalarEvolution &SE)
Check if VPV is an untruncated wide induction, either before or after the increment.
static VPRegionBlock * createReplicateRegion(VPReplicateRecipe *PredRecipe, VPlan &Plan)
static VPBasicBlock * getPredicatedThenBlock(VPRegionBlock *R)
If R is a triangle region, return the 'then' block of the triangle.
static void simplifyBlends(VPlan &Plan)
Normalize and simplify VPBlendRecipes.
static bool isAlreadyNarrow(VPValue *VPV)
Returns true if VPValue is a narrow VPValue.
static bool optimizeVectorInductionWidthForTCAndVFUF(VPlan &Plan, ElementCount BestVF, unsigned BestUF)
Optimize the width of vector induction variables in Plan based on a known constant Trip Count,...
VPValue * getPredicatedMask(VPRegionBlock *R)
If R is a region with a VPBranchOnMaskRecipe in the entry block, return the mask.
static VPExpressionRecipe * tryToMatchAndCreateMulAccumulateReduction(VPReductionRecipe *Red, VPCostContext &Ctx, VFRange &Range)
This function tries convert extended in-loop reductions to VPExpressionRecipe and clamp the Range if ...
static void expandVPWidenIntOrFpInduction(VPWidenIntOrFpInductionRecipe *WidenIVR, VPTypeAnalysis &TypeInfo)
Expand a VPWidenIntOrFpInduction into executable recipes, for the initial value, phi and backedge val...
static VPSingleDefRecipe * findHeaderMask(VPlan &Plan)
Collect the header mask with the pattern: (ICMP_ULE, WideCanonicalIV, backedge-taken-count) TODO: Int...
static VPRecipeBase * optimizeMaskToEVL(VPValue *HeaderMask, VPRecipeBase &CurRecipe, VPTypeAnalysis &TypeInfo, VPValue &AllOneMask, VPValue &EVL)
Try to optimize a CurRecipe masked by HeaderMask to a corresponding EVL-based recipe without the head...
static void removeRedundantCanonicalIVs(VPlan &Plan)
Try to replace VPWidenCanonicalIVRecipes with a widened canonical IV recipe, if it exists.
static bool canNarrowLoad(VPWidenRecipe *WideMember0, unsigned OpIdx, VPValue *OpV, unsigned Idx)
Returns true if V is VPWidenLoadRecipe or VPInterleaveRecipe that can be converted to a narrower reci...
static void narrowToSingleScalarRecipes(VPlan &Plan)
This file provides utility VPlan to VPlan transformations.
This file declares the class VPlanVerifier, which contains utility functions to check the consistency...
This file contains the declarations of the Vectorization Plan base classes:
Value * RHS
Value * LHS
BinaryOperator * Mul
static const uint32_t IV[8]
Definition blake3_impl.h:83
Class for arbitrary precision integers.
Definition APInt.h:78
LLVM_ABI APInt zext(unsigned width) const
Zero extend to a new width.
Definition APInt.cpp:1012
unsigned getActiveBits() const
Compute the number of active bits in the value.
Definition APInt.h:1512
unsigned getBitWidth() const
Return the number of bits in the APInt.
Definition APInt.h:1488
LLVM_ABI APInt sext(unsigned width) const
Sign extend to a new width.
Definition APInt.cpp:985
ArrayRef - Represent a constant reference to an array (0 or more elements consecutively in memory),...
Definition ArrayRef.h:41
LLVM Basic Block Representation.
Definition BasicBlock.h:62
const Function * getParent() const
Return the enclosing method, or null if none.
Definition BasicBlock.h:213
LLVM_ABI const DataLayout & getDataLayout() const
Get the data layout of the module this basic block belongs to.
const Instruction * getTerminator() const LLVM_READONLY
Returns the terminator instruction if the block is well formed or null if the block is not well forme...
Definition BasicBlock.h:233
This class represents a function call, abstracting a target machine's calling convention.
@ ICMP_ULT
unsigned less than
Definition InstrTypes.h:703
@ ICMP_ULE
unsigned less or equal
Definition InstrTypes.h:704
Predicate getInversePredicate() const
For example, EQ -> NE, UGT -> ULE, SLT -> SGE, OEQ -> UNE, UGT -> OLE, OLT -> UGE,...
Definition InstrTypes.h:791
An abstraction over a floating-point predicate, and a pack of an integer predicate with samesign info...
static ConstantInt * getSigned(IntegerType *Ty, int64_t V)
Return a ConstantInt with the specified value for the specified type.
Definition Constants.h:131
static LLVM_ABI ConstantInt * getFalse(LLVMContext &Context)
static LLVM_ABI Constant * getAllOnesValue(Type *Ty)
static LLVM_ABI Constant * getNullValue(Type *Ty)
Constructor to create a '0' constant of arbitrary type.
A parsed version of the target data layout string in and methods for querying it.
Definition DataLayout.h:63
A debug info location.
Definition DebugLoc.h:124
static DebugLoc getCompilerGenerated()
Definition DebugLoc.h:163
static DebugLoc getUnknown()
Definition DebugLoc.h:162
ValueT lookup(const_arg_type_t< KeyT > Val) const
lookup - Return the entry for the specified key, or a default constructed value if no such entry exis...
Definition DenseMap.h:187
std::pair< iterator, bool > try_emplace(KeyT &&Key, Ts &&...Args)
Definition DenseMap.h:229
bool dominates(const DomTreeNodeBase< NodeT > *A, const DomTreeNodeBase< NodeT > *B) const
dominates - Returns true iff A dominates B.
void recalculate(ParentType &Func)
recalculate - compute a dominator tree for the given function
constexpr bool isVector() const
One or more elements.
Definition TypeSize.h:324
static constexpr ElementCount getScalable(ScalarTy MinVal)
Definition TypeSize.h:312
Utility class for floating point operations which can have information about relaxed accuracy require...
Definition Operator.h:200
Represents flags for the getelementptr instruction/expression.
GEPNoWrapFlags withoutNoUnsignedWrap() const
static GEPNoWrapFlags none()
an instruction for type-safe pointer arithmetic to access elements of arrays and structs
A struct for saving information about induction variables.
InductionKind
This enum represents the kinds of inductions that we support.
@ IK_PtrInduction
Pointer induction var. Step = C.
@ IK_IntInduction
Integer induction variable. Step = C.
InstSimplifyFolder - Use InstructionSimplify to fold operations to existing values.
bool isCast() const
bool isBinaryOp() const
LLVM_ABI const DataLayout & getDataLayout() const
Get the data layout of the module this instruction belongs to.
static LLVM_ABI IntegerType * get(LLVMContext &C, unsigned NumBits)
This static method is the primary way of constructing an IntegerType.
Definition Type.cpp:319
The group of interleaved loads/stores sharing the same stride and close to each other.
InstTy * getMember(uint32_t Index) const
Get the member with the given index Index.
uint32_t getNumMembers() const
This is an important class for using LLVM in a threaded context.
Definition LLVMContext.h:68
An instruction for reading from memory.
static bool getDecisionAndClampRange(const std::function< bool(ElementCount)> &Predicate, VFRange &Range)
Test a Predicate on a Range of VF's.
Definition VPlan.cpp:1576
LLVM_ABI MDNode * createBranchWeights(uint32_t TrueWeight, uint32_t FalseWeight, bool IsExpected=false)
Return metadata containing two branch weights.
Definition MDBuilder.cpp:38
Metadata node.
Definition Metadata.h:1077
This class implements a map that also provides access to all stored values in a deterministic order.
Definition MapVector.h:36
ValueT lookup(const KeyT &Key) const
Definition MapVector.h:99
An interface layer with SCEV used to manage how we see SCEV expressions for values in the context of ...
ScalarEvolution * getSE() const
Returns the ScalarEvolution analysis used.
LLVM_ABI const SCEV * getSCEV(Value *V)
Returns the SCEV expression of V, in the context of the current SCEV predicate.
static LLVM_ABI unsigned getOpcode(RecurKind Kind)
Returns the opcode corresponding to the RecurrenceKind.
RegionT * getParent() const
Get the parent of the Region.
Definition RegionInfo.h:362
This class uses information about analyze scalars to rewrite expressions in canonical form.
LLVM_ABI Value * expandCodeFor(const SCEV *SH, Type *Ty, BasicBlock::iterator I)
Insert code to directly compute the specified SCEV expression into the program.
This class represents an analyzed expression in the program.
LLVM_ABI bool isZero() const
Return true if the expression is a constant zero.
LLVM_ABI Type * getType() const
Return the LLVM type of this SCEV expression.
The main scalar evolution driver.
const DataLayout & getDataLayout() const
Return the DataLayout associated with the module this SCEV instance is operating on.
LLVM_ABI const SCEV * getNegativeSCEV(const SCEV *V, SCEV::NoWrapFlags Flags=SCEV::FlagAnyWrap)
Return the SCEV object corresponding to -V.
LLVM_ABI const SCEV * getSCEV(Value *V)
Return a SCEV expression for the full generality of the specified expression.
LLVM_ABI const SCEV * getUDivExpr(const SCEV *LHS, const SCEV *RHS)
Get a canonical unsigned division expression, or something simpler if possible.
LLVM_ABI const SCEV * getElementCount(Type *Ty, ElementCount EC, SCEV::NoWrapFlags Flags=SCEV::FlagAnyWrap)
LLVM_ABI const SCEV * getMulExpr(SmallVectorImpl< const SCEV * > &Ops, SCEV::NoWrapFlags Flags=SCEV::FlagAnyWrap, unsigned Depth=0)
Get a canonical multiply expression, or something simpler if possible.
LLVM_ABI bool isKnownPredicate(CmpPredicate Pred, const SCEV *LHS, const SCEV *RHS)
Test if the given expression is known to satisfy the condition described by Pred, LHS,...
This class represents the LLVM 'select' instruction.
A vector that has set insertion semantics.
Definition SetVector.h:59
size_type size() const
Determine the number of elements in the SetVector.
Definition SetVector.h:104
bool insert(const value_type &X)
Insert a new element into the SetVector.
Definition SetVector.h:168
size_type size() const
Definition SmallPtrSet.h:99
A templated base class for SmallPtrSet which provides the typesafe interface that is common across al...
std::pair< iterator, bool > insert(PtrType Ptr)
Inserts Ptr if and only if there is no element in the container equal to Ptr.
iterator begin() const
bool contains(ConstPtrType Ptr) const
SmallPtrSet - This class implements a set which is optimized for holding SmallSize or less elements.
This class consists of common code factored out of the SmallVector class to reduce code duplication b...
void append(ItTy in_start, ItTy in_end)
Add the specified range to the end of the SmallVector.
void push_back(const T &Elt)
This is a 'vector' (really, a variable-sized array), optimized for the case when the array is small.
An instruction for storing to memory.
Provides information about what library functions are available for the current target.
TargetCostKind
The kind of cost model.
@ TCK_RecipThroughput
Reciprocal throughput.
Twine - A lightweight data structure for efficiently representing the concatenation of temporary valu...
Definition Twine.h:82
This class implements a switch-like dispatch statement for a value of 'T' using dyn_cast functionalit...
Definition TypeSwitch.h:87
TypeSwitch< T, ResultT > & Case(CallableT &&caseFn)
Add a case on the given type.
Definition TypeSwitch.h:96
The instances of the Type class are immutable: once they are created, they are never changed.
Definition Type.h:45
static LLVM_ABI IntegerType * getInt64Ty(LLVMContext &C)
Definition Type.cpp:298
static LLVM_ABI IntegerType * getInt32Ty(LLVMContext &C)
Definition Type.cpp:297
bool isPointerTy() const
True if this is an instance of PointerType.
Definition Type.h:267
static LLVM_ABI IntegerType * getInt8Ty(LLVMContext &C)
Definition Type.cpp:295
bool isStructTy() const
True if this is an instance of StructType.
Definition Type.h:261
LLVM_ABI unsigned getScalarSizeInBits() const LLVM_READONLY
If this is a vector type, return the getPrimitiveSizeInBits value for the element type.
Definition Type.cpp:231
static LLVM_ABI IntegerType * getInt1Ty(LLVMContext &C)
Definition Type.cpp:294
bool isFloatingPointTy() const
Return true if this is one of the floating-point types.
Definition Type.h:184
bool isIntegerTy() const
True if this is an instance of IntegerType.
Definition Type.h:240
op_range operands()
Definition User.h:292
A recipe for generating the active lane mask for the vector loop that is used to predicate the vector...
Definition VPlan.h:3468
VPBasicBlock serves as the leaf of the Hierarchical Control-Flow Graph.
Definition VPlan.h:3755
void appendRecipe(VPRecipeBase *Recipe)
Augment the existing recipes of a VPBasicBlock with an additional Recipe as the last recipe.
Definition VPlan.h:3830
RecipeListTy::iterator iterator
Instruction iterators...
Definition VPlan.h:3782
iterator end()
Definition VPlan.h:3792
iterator begin()
Recipe iterator methods.
Definition VPlan.h:3790
iterator_range< iterator > phis()
Returns an iterator range over the PHI-like recipes in the block.
Definition VPlan.h:3843
iterator getFirstNonPhi()
Return the position of the first non-phi node recipe in the block.
Definition VPlan.cpp:246
VPRegionBlock * getEnclosingLoopRegion()
Definition VPlan.cpp:619
VPBasicBlock * splitAt(iterator SplitAt)
Split current block at SplitAt by inserting a new block between the current block and its successors ...
Definition VPlan.cpp:591
VPRecipeBase * getTerminator()
If the block has multiple successors, return the branch recipe terminating the block.
Definition VPlan.cpp:664
const VPRecipeBase & back() const
Definition VPlan.h:3804
A recipe for vectorizing a phi-node as a sequence of mask-based select instructions.
Definition VPlan.h:2394
VPValue * getMask(unsigned Idx) const
Return mask number Idx.
Definition VPlan.h:2428
unsigned getNumIncomingValues() const
Return the number of incoming values, taking into account when normalized the first incoming value wi...
Definition VPlan.h:2418
void setMask(unsigned Idx, VPValue *V)
Set mask number Idx to V.
Definition VPlan.h:2434
bool isNormalized() const
A normalized blend is one that has an odd number of operands, whereby the first operand does not have...
Definition VPlan.h:2414
VPBlockBase is the building block of the Hierarchical Control-Flow Graph.
Definition VPlan.h:81
VPRegionBlock * getParent()
Definition VPlan.h:173
const VPBasicBlock * getExitingBasicBlock() const
Definition VPlan.cpp:190
void swapSuccessors()
Swap successors of the block. The block must have exactly 2 successors.
Definition VPlan.h:322
size_t getNumPredecessors() const
Definition VPlan.h:220
const VPBlocksTy & getPredecessors() const
Definition VPlan.h:204
VPlan * getPlan()
Definition VPlan.cpp:165
VPBlockBase * getSinglePredecessor() const
Definition VPlan.h:215
const VPBasicBlock * getEntryBasicBlock() const
Definition VPlan.cpp:170
VPBlockBase * getSingleHierarchicalPredecessor()
Definition VPlan.h:264
VPBlockBase * getSingleSuccessor() const
Definition VPlan.h:209
const VPBlocksTy & getSuccessors() const
Definition VPlan.h:198
static auto blocksOnly(const T &Range)
Return an iterator range over Range which only includes BlockTy blocks.
Definition VPlanUtils.h:228
static void insertOnEdge(VPBlockBase *From, VPBlockBase *To, VPBlockBase *BlockPtr)
Inserts BlockPtr on the edge between From and To.
Definition VPlanUtils.h:249
static void insertTwoBlocksAfter(VPBlockBase *IfTrue, VPBlockBase *IfFalse, VPBlockBase *BlockPtr)
Insert disconnected VPBlockBases IfTrue and IfFalse after BlockPtr.
Definition VPlanUtils.h:168
static void connectBlocks(VPBlockBase *From, VPBlockBase *To, unsigned PredIdx=-1u, unsigned SuccIdx=-1u)
Connect VPBlockBases From and To bi-directionally.
Definition VPlanUtils.h:187
static void disconnectBlocks(VPBlockBase *From, VPBlockBase *To)
Disconnect VPBlockBases From and To bi-directionally.
Definition VPlanUtils.h:206
A recipe for generating conditional branches on the bits of a mask.
Definition VPlan.h:2925
RAII object that stores the current insertion point and restores it when the object is destroyed.
VPlan-based builder utility analogous to IRBuilder.
VPValue * createElementCount(Type *Ty, ElementCount EC)
static VPBuilder getToInsertAfter(VPRecipeBase *R)
Create a VPBuilder to insert after R.
VPPhi * createScalarPhi(ArrayRef< VPValue * > IncomingValues, DebugLoc DL, const Twine &Name="")
VPInstruction * createNaryOp(unsigned Opcode, ArrayRef< VPValue * > Operands, Instruction *Inst=nullptr, const Twine &Name="")
Create an N-ary operation with Opcode, Operands and set Inst as its underlying Instruction.
void setInsertPoint(VPBasicBlock *TheBB)
This specifies that created VPInstructions should be appended to the end of the specified block.
Canonical scalar induction phi of the vector loop.
Definition VPlan.h:3411
Type * getScalarType() const
Returns the scalar type of the induction.
Definition VPlan.h:3438
unsigned getNumDefinedValues() const
Returns the number of values defined by the VPDef.
Definition VPlanValue.h:424
ArrayRef< VPValue * > definedValues()
Returns an ArrayRef of the values defined by the VPDef.
Definition VPlanValue.h:419
VPValue * getVPSingleValue()
Returns the only VPValue defined by the VPDef.
Definition VPlanValue.h:397
VPValue * getVPValue(unsigned I)
Returns the VPValue with index I defined by the VPDef.
Definition VPlanValue.h:409
Template specialization of the standard LLVM dominator tree utility for VPBlockBases.
bool properlyDominates(const VPRecipeBase *A, const VPRecipeBase *B)
A recipe for generating the phi node for the current index of elements, adjusted in accordance with E...
Definition VPlan.h:3499
A recipe to combine multiple recipes into a single 'expression' recipe, which should be considered a ...
Definition VPlan.h:2970
A pure virtual base class for all recipes modeling header phis, including phis for first order recurr...
Definition VPlan.h:1968
virtual VPValue * getBackedgeValue()
Returns the incoming value from the loop backedge.
Definition VPlan.h:2016
VPValue * getStartValue()
Returns the start value of the phi, if one is set.
Definition VPlan.h:2005
A special type of VPBasicBlock that wraps an existing IR basic block.
Definition VPlan.h:3908
BasicBlock * getIRBasicBlock() const
Definition VPlan.h:3932
Class to record and manage LLVM IR flags.
Definition VPlan.h:600
static LLVM_ABI_FOR_TEST VPIRInstruction * create(Instruction &I)
Create a new VPIRPhi for \I , if it is a PHINode, otherwise create a VPIRInstruction.
Helper to manage IR metadata for recipes.
Definition VPlan.h:939
void intersect(const VPIRMetadata &MD)
Intersect this VPIRMetada object with MD, keeping only metadata nodes that are common to both.
This is a concrete Recipe that models a single VPlan-level instruction.
Definition VPlan.h:980
@ ExtractLane
Extracts a single lane (first operand) from a set of vector operands.
Definition VPlan.h:1057
@ FirstOrderRecurrenceSplice
Definition VPlan.h:986
@ BuildVector
Creates a fixed-width vector containing all operands.
Definition VPlan.h:1010
@ BuildStructVector
Given operands of (the same) struct type, creates a struct of fixed- width vectors each containing a ...
Definition VPlan.h:1007
@ CanonicalIVIncrementForPart
Definition VPlan.h:1000
@ CalculateTripCountMinusVF
Definition VPlan.h:998
const InterleaveGroup< Instruction > * getInterleaveGroup() const
Definition VPlan.h:2535
ArrayRef< VPValue * > getStoredValues() const
Return the VPValues stored by this interleave group.
Definition VPlan.h:2556
A recipe for interleaved memory operations with vector-predication intrinsics.
Definition VPlan.h:2608
VPInterleaveRecipe is a recipe for transforming an interleave group of load or stores into one wide l...
Definition VPlan.h:2567
VPPredInstPHIRecipe is a recipe for generating the phi nodes needed when control converges back from ...
Definition VPlan.h:3082
VPRecipeBase is a base class modeling a sequence of one or more output IR instructions.
Definition VPlan.h:394
bool mayReadFromMemory() const
Returns true if the recipe may read from memory.
bool mayReadOrWriteMemory() const
Returns true if the recipe may read from or write to memory.
Definition VPlan.h:477
bool mayHaveSideEffects() const
Returns true if the recipe may have side-effects.
VPBasicBlock * getParent()
Definition VPlan.h:415
DebugLoc getDebugLoc() const
Returns the debug location of the recipe.
Definition VPlan.h:482
void moveBefore(VPBasicBlock &BB, iplist< VPRecipeBase >::iterator I)
Unlink this recipe and insert into BB before I.
void insertBefore(VPRecipeBase *InsertPos)
Insert an unlinked recipe into a basic block immediately before the specified recipe.
void insertAfter(VPRecipeBase *InsertPos)
Insert an unlinked Recipe into a basic block immediately after the specified Recipe.
iplist< VPRecipeBase >::iterator eraseFromParent()
This method unlinks 'this' from the containing basic block and deletes it.
Helper class to create VPRecipies from IR instructions.
VPRecipeBase * getRecipe(Instruction *I)
Return the recipe created for given ingredient.
A recipe to represent inloop reduction operations with vector-predication intrinsics,...
Definition VPlan.h:2803
A recipe to represent inloop reduction operations, performing a reduction on a vector operand into a ...
Definition VPlan.h:2657
VPRegionBlock represents a collection of VPBasicBlocks and VPRegionBlocks which form a Single-Entry-S...
Definition VPlan.h:3943
const VPBlockBase * getEntry() const
Definition VPlan.h:3979
void setExiting(VPBlockBase *ExitingBlock)
Set ExitingBlock as the exiting VPBlockBase of this VPRegionBlock.
Definition VPlan.h:3996
const VPBlockBase * getExiting() const
Definition VPlan.h:3991
VPBasicBlock * getPreheaderVPBB()
Returns the pre-header VPBasicBlock of the loop region.
Definition VPlan.h:4004
VPReplicateRecipe replicates a given instruction producing multiple scalar copies of the original sca...
Definition VPlan.h:2847
bool isSingleScalar() const
Definition VPlan.h:2892
VPValue * getMask()
Return the mask of a predicated VPReplicateRecipe.
Definition VPlan.h:2916
A recipe for handling phi nodes of integer and floating-point inductions, producing their scalar valu...
Definition VPlan.h:3645
VPSingleDef is a base class for recipes for modeling a sequence of one or more output IR that define ...
Definition VPlan.h:521
Instruction * getUnderlyingInstr()
Returns the underlying instruction.
Definition VPlan.h:586
virtual VPSingleDefRecipe * clone() override=0
Clone the current recipe.
An analysis for type-inference for VPValues.
LLVMContext & getContext()
Return the LLVMContext used by the analysis.
Type * inferScalarType(const VPValue *V)
Infer the type of V. Returns the scalar type of V.
This class augments VPValue with operands which provide the inverse def-use edges from VPValue's user...
Definition VPlanValue.h:199
operand_range operands()
Definition VPlanValue.h:267
void setOperand(unsigned I, VPValue *New)
Definition VPlanValue.h:243
operand_iterator op_end()
Definition VPlanValue.h:265
operand_iterator op_begin()
Definition VPlanValue.h:263
VPValue * getOperand(unsigned N) const
Definition VPlanValue.h:238
void addOperand(VPValue *Operand)
Definition VPlanValue.h:232
VPRecipeBase * getDefiningRecipe()
Returns the recipe defining this VPValue or nullptr if it is not defined by a recipe,...
Definition VPlan.cpp:135
Value * getLiveInIRValue() const
Returns the underlying IR value, if this VPValue is defined outside the scope of VPlan.
Definition VPlanValue.h:176
Value * getUnderlyingValue() const
Return the underlying Value attached to this VPValue.
Definition VPlanValue.h:85
void setUnderlyingValue(Value *Val)
Definition VPlanValue.h:186
void replaceAllUsesWith(VPValue *New)
Definition VPlan.cpp:1412
unsigned getNumUsers() const
Definition VPlanValue.h:113
bool isLiveIn() const
Returns true if this VPValue is a live-in, i.e. defined outside the VPlan.
Definition VPlanValue.h:171
void replaceUsesWithIf(VPValue *New, llvm::function_ref< bool(VPUser &U, unsigned Idx)> ShouldReplace)
Go through the uses list for this VPValue and make each use point to New if the callback ShouldReplac...
Definition VPlan.cpp:1416
user_range users()
Definition VPlanValue.h:134
A recipe to compute a pointer to the last element of each part of a widened memory access for widened...
Definition VPlan.h:1832
VPVectorEndPointerRecipe * clone() override
Clone the current recipe.
Definition VPlan.h:1876
A Recipe for widening the canonical induction variable of the vector loop.
Definition VPlan.h:3540
VPWidenCastRecipe is a recipe to create vector cast instructions.
Definition VPlan.h:1479
A recipe for handling GEP instructions.
Definition VPlan.h:1765
Base class for widened induction (VPWidenIntOrFpInductionRecipe and VPWidenPointerInductionRecipe),...
Definition VPlan.h:2033
PHINode * getPHINode() const
Definition VPlan.h:2075
VPValue * getStepValue()
Returns the step value of the induction.
Definition VPlan.h:2061
const InductionDescriptor & getInductionDescriptor() const
Returns the induction descriptor for the recipe.
Definition VPlan.h:2078
A recipe for handling phi nodes of integer and floating-point inductions, producing their vector valu...
Definition VPlan.h:2108
VPValue * getLastUnrolledPartOperand()
Returns the VPValue representing the value of this induction at the last unrolled part,...
Definition VPlan.h:2189
A recipe for widening vector intrinsics.
Definition VPlan.h:1536
A common base class for widening memory operations.
Definition VPlan.h:3124
VPValue * getMask() const
Return the mask used by this recipe.
Definition VPlan.h:3186
VPValue * getAddr() const
Return the address accessed by this recipe.
Definition VPlan.h:3179
A recipe for widened phis.
Definition VPlan.h:2244
VPWidenRecipe is a recipe for producing a widened instruction using the opcode and operands of the re...
Definition VPlan.h:1436
VPlan models a candidate for vectorization, encoding various decisions take to produce efficient outp...
Definition VPlan.h:4046
bool hasVF(ElementCount VF) const
Definition VPlan.h:4255
LLVMContext & getContext() const
Definition VPlan.h:4243
VPBasicBlock * getEntry()
Definition VPlan.h:4145
VPRegionBlock * createVPRegionBlock(VPBlockBase *Entry, VPBlockBase *Exiting, const std::string &Name="", bool IsReplicator=false)
Create a new VPRegionBlock with Entry, Exiting and Name.
Definition VPlan.h:4386
VPValue & getVectorTripCount()
The vector trip count.
Definition VPlan.h:4235
bool hasScalableVF() const
Definition VPlan.h:4256
VPValue & getVFxUF()
Returns VF * UF of the vector loop region.
Definition VPlan.h:4241
VPValue & getVF()
Returns the VF of the vector loop region.
Definition VPlan.h:4238
VPValue * getTripCount() const
The trip count of the original loop.
Definition VPlan.h:4207
VPValue * getTrue()
Return a VPValue wrapping i1 true.
Definition VPlan.h:4312
VPValue * getOrCreateBackedgeTakenCount()
The backedge taken count of the original loop.
Definition VPlan.h:4228
unsigned getUF() const
Definition VPlan.h:4275
bool hasUF(unsigned UF) const
Definition VPlan.h:4273
ArrayRef< VPIRBasicBlock * > getExitBlocks() const
Return an ArrayRef containing VPIRBasicBlocks wrapping the exit blocks of the original scalar loop.
Definition VPlan.h:4197
void setVF(ElementCount VF)
Definition VPlan.h:4249
bool isUnrolled() const
Returns true if the VPlan already has been unrolled, i.e.
Definition VPlan.h:4288
LLVM_ABI_FOR_TEST VPRegionBlock * getVectorLoopRegion()
Returns the VPRegionBlock of the vector loop.
Definition VPlan.cpp:1046
void resetTripCount(VPValue *NewTripCount)
Resets the trip count for the VPlan.
Definition VPlan.h:4221
VPBasicBlock * getMiddleBlock()
Returns the 'middle' block of the plan, that is the block that selects whether to execute the scalar ...
Definition VPlan.h:4170
VPBasicBlock * createVPBasicBlock(const Twine &Name, VPRecipeBase *Recipe=nullptr)
Create a new VPBasicBlock with Name and containing Recipe if present.
Definition VPlan.h:4376
VPValue * getOrAddLiveIn(Value *V)
Gets the live-in VPValue for V or adds a new live-in (if none exists yet) for V.
Definition VPlan.h:4297
bool hasScalarVFOnly() const
Definition VPlan.h:4266
VPBasicBlock * getScalarPreheader() const
Return the VPBasicBlock for the preheader of the scalar loop.
Definition VPlan.h:4188
ArrayRef< VPValue * > getLiveIns() const
Return the list of live-in VPValues available in the VPlan.
Definition VPlan.h:4327
VPCanonicalIVPHIRecipe * getCanonicalIV()
Returns the canonical induction recipe of the vector loop.
Definition VPlan.h:4351
VPIRBasicBlock * getScalarHeader() const
Return the VPIRBasicBlock wrapping the header of the scalar loop.
Definition VPlan.h:4193
VPValue * getLiveIn(Value *V) const
Return the live-in VPValue for V, if there is one or nullptr otherwise.
Definition VPlan.h:4324
VPBasicBlock * getVectorPreheader()
Returns the preheader of the vector loop region, if one exists, or null otherwise.
Definition VPlan.h:4150
void setUF(unsigned UF)
Definition VPlan.h:4280
bool hasScalarTail() const
Returns true if the scalar tail may execute after the vector loop.
Definition VPlan.h:4428
LLVM Value Representation.
Definition Value.h:75
Type * getType() const
All values are typed, get the type of this value.
Definition Value.h:256
iterator_range< user_iterator > users()
Definition Value.h:426
bool hasName() const
Definition Value.h:262
LLVM_ABI StringRef getName() const
Return a constant reference to the value's name.
Definition Value.cpp:322
constexpr bool isScalable() const
Returns whether the quantity is scaled by a runtime quantity (vscale).
Definition TypeSize.h:169
constexpr LeafTy multiplyCoefficientBy(ScalarTy RHS) const
Definition TypeSize.h:256
constexpr bool isFixed() const
Returns true if the quantity is not scaled by vscale.
Definition TypeSize.h:172
constexpr ScalarTy getKnownMinValue() const
Returns the minimum value this quantity can represent.
Definition TypeSize.h:166
An efficient, type-erasing, non-owning reference to a callable.
const ParentTy * getParent() const
Definition ilist_node.h:34
self_iterator getIterator()
Definition ilist_node.h:130
IteratorT end() const
Changed
#define llvm_unreachable(msg)
Marks that the current location is not supposed to be reachable.
LLVM_ABI APInt RoundingUDiv(const APInt &A, const APInt &B, APInt::Rounding RM)
Return A unsign-divided by B, rounded by the given rounding mode.
Definition APInt.cpp:2763
unsigned ID
LLVM IR allows to use arbitrary numbers as calling convention identifiers.
Definition CallingConv.h:24
@ C
The default llvm calling convention, compatible with C.
Definition CallingConv.h:34
SpecificConstantMatch m_ZeroInt()
Convenience matchers for specific integer values.
BinaryOp_match< SrcTy, SpecificConstantMatch, TargetOpcode::G_XOR, true > m_Not(const SrcTy &&Src)
Matches a register not-ed by a G_XOR.
cst_pred_ty< is_all_ones > m_AllOnes()
Match an integer or vector with all bits set.
CastInst_match< OpTy, TruncInst > m_Trunc(const OpTy &Op)
Matches Trunc.
LogicalOp_match< LHS, RHS, Instruction::And > m_LogicalAnd(const LHS &L, const RHS &R)
Matches L && R either in the form of L & R or L ?
match_combine_or< CastInst_match< OpTy, ZExtInst >, OpTy > m_ZExtOrSelf(const OpTy &Op)
bool match(Val *V, const Pattern &P)
specificval_ty m_Specific(const Value *V)
Match if we have a specific specified value.
cst_pred_ty< is_one > m_One()
Match an integer 1 or a vector with all elements equal to 1.
IntrinsicID_match m_Intrinsic()
Match intrinsic calls like this: m_Intrinsic<Intrinsic::fabs>(m_Value(X))
ThreeOps_match< Cond, LHS, RHS, Instruction::Select > m_Select(const Cond &C, const LHS &L, const RHS &R)
Matches SelectInst.
BinaryOp_match< LHS, RHS, Instruction::Mul > m_Mul(const LHS &L, const RHS &R)
deferredval_ty< Value > m_Deferred(Value *const &V)
Like m_Specific(), but works if the specific value to match is determined as part of the same match()...
SpecificCmpClass_match< LHS, RHS, ICmpInst > m_SpecificICmp(CmpPredicate MatchPred, const LHS &L, const RHS &R)
class_match< CmpInst > m_Cmp()
Matches any compare instruction and ignore it.
BinaryOp_match< LHS, RHS, Instruction::Add, true > m_c_Add(const LHS &L, const RHS &R)
Matches a Add with LHS and RHS in either order.
CmpClass_match< LHS, RHS, ICmpInst > m_ICmp(CmpPredicate &Pred, const LHS &L, const RHS &R)
match_combine_or< CastInst_match< OpTy, ZExtInst >, CastInst_match< OpTy, SExtInst > > m_ZExtOrSExt(const OpTy &Op)
auto m_LogicalAnd()
Matches L && R where L and R are arbitrary values.
CastInst_match< OpTy, SExtInst > m_SExt(const OpTy &Op)
Matches SExt.
BinaryOp_match< LHS, RHS, Instruction::Mul, true > m_c_Mul(const LHS &L, const RHS &R)
Matches a Mul with LHS and RHS in either order.
MatchFunctor< Val, Pattern > match_fn(const Pattern &P)
A match functor that can be used as a UnaryPredicate in functional algorithms like all_of.
BinaryOp_match< LHS, RHS, Instruction::Sub > m_Sub(const LHS &L, const RHS &R)
match_combine_or< LTy, RTy > m_CombineOr(const LTy &L, const RTy &R)
Combine two pattern matchers matching L || R.
AllRecipe_commutative_match< Instruction::And, Op0_t, Op1_t > m_c_BinaryAnd(const Op0_t &Op0, const Op1_t &Op1)
Match a binary AND operation.
AllRecipe_match< Instruction::Or, Op0_t, Op1_t > m_BinaryOr(const Op0_t &Op0, const Op1_t &Op1)
Match a binary OR operation.
AllRecipe_commutative_match< Opcode, Op0_t, Op1_t > m_c_Binary(const Op0_t &Op0, const Op1_t &Op1)
AllRecipe_commutative_match< Instruction::Or, Op0_t, Op1_t > m_c_BinaryOr(const Op0_t &Op0, const Op1_t &Op1)
GEPLikeRecipe_match< Op0_t, Op1_t > m_GetElementPtr(const Op0_t &Op0, const Op1_t &Op1)
VPInstruction_match< VPInstruction::ExtractLastElement, Op0_t > m_ExtractLastElement(const Op0_t &Op0)
AllRecipe_match< Opcode, Op0_t, Op1_t > m_Binary(const Op0_t &Op0, const Op1_t &Op1)
specific_intval< 1 > m_False()
VPDerivedIV_match< Op0_t, Op1_t, Op2_t > m_DerivedIV(const Op0_t &Op0, const Op1_t &Op1, const Op2_t &Op2)
VPInstruction_match< VPInstruction::ActiveLaneMask, Op0_t, Op1_t, Op2_t > m_ActiveLaneMask(const Op0_t &Op0, const Op1_t &Op1, const Op2_t &Op2)
VPInstruction_match< VPInstruction::BranchOnCount, Op0_t, Op1_t > m_BranchOnCount(const Op0_t &Op0, const Op1_t &Op1)
specific_intval< 1 > m_True()
VPInstruction_match< VPInstruction::Broadcast, Op0_t > m_Broadcast(const Op0_t &Op0)
class_match< VPValue > m_VPValue()
Match an arbitrary VPValue and ignore it.
VPInstruction_match< VPInstruction::ExplicitVectorLength, Op0_t > m_EVL(const Op0_t &Op0)
VPInstruction_match< VPInstruction::BuildVector > m_BuildVector()
BuildVector is matches only its opcode, w/o matching its operands as the number of operands is not fi...
VPInstruction_match< VPInstruction::BranchOnCond, Op0_t > m_BranchOnCond(const Op0_t &Op0)
bind_ty< VPInstruction > m_VPInstruction(VPInstruction *&V)
Match a VPInstruction, capturing if we match.
initializer< Ty > init(const Ty &Val)
NodeAddr< DefNode * > Def
Definition RDFGraph.h:384
bool isSingleScalar(const VPValue *VPV)
Returns true if VPV is a single scalar, either because it produces the same value for all lanes or on...
Definition VPlanUtils.h:44
VPValue * getOrCreateVPValueForSCEVExpr(VPlan &Plan, const SCEV *Expr)
Get or create a VPValue that corresponds to the expansion of Expr.
const SCEV * getSCEVExprForVPValue(VPValue *V, ScalarEvolution &SE)
Return the SCEV expression for V.
bool onlyFirstLaneUsed(const VPValue *Def)
Returns true if only the first lane of Def is used.
bool isHeaderMask(const VPValue *V, VPlan &Plan)
Return true if V is a header mask in Plan.
bool onlyScalarValuesUsed(const VPValue *Def)
Returns true if only scalar values of Def are used by all users.
This is an optimization pass for GlobalISel generic memory operations.
auto drop_begin(T &&RangeOrContainer, size_t N=1)
Return a range covering RangeOrContainer with the first N elements excluded.
Definition STLExtras.h:318
@ Offset
Definition DWP.cpp:477
bool all_of(R &&range, UnaryPredicate P)
Provide wrappers to std::all_of which take ranges instead of having to pass begin/end explicitly.
Definition STLExtras.h:1705
LLVM_ABI Intrinsic::ID getVectorIntrinsicIDForCall(const CallInst *CI, const TargetLibraryInfo *TLI)
Returns intrinsic ID for call.
auto enumerate(FirstRange &&First, RestRanges &&...Rest)
Given two or more input ranges, returns a new range whose values are tuples (A, B,...
Definition STLExtras.h:2452
decltype(auto) dyn_cast(const From &Val)
dyn_cast<X> - Return the argument parameter cast to the specified type.
Definition Casting.h:649
const Value * getLoadStorePointerOperand(const Value *V)
A helper function that returns the pointer operand of a load or store instruction.
constexpr from_range_t from_range
auto dyn_cast_if_present(const Y &Val)
dyn_cast_if_present<X> - Functionally identical to dyn_cast, except that a null (or none in the case ...
Definition Casting.h:738
iterator_range< T > make_range(T x, T y)
Convenience function for iterating over sub-ranges.
void append_range(Container &C, Range &&R)
Wrapper function to append range R to container C.
Definition STLExtras.h:2116
iterator_range< early_inc_iterator_impl< detail::IterOfRange< RangeT > > > make_early_inc_range(RangeT &&Range)
Make a range that does early increment to allow mutation of the underlying range without disrupting i...
Definition STLExtras.h:634
auto cast_or_null(const Y &Val)
Definition Casting.h:720
iterator_range< df_iterator< VPBlockShallowTraversalWrapper< VPBlockBase * > > > vp_depth_first_shallow(VPBlockBase *G)
Returns an iterator range to traverse the graph starting at G in depth-first order.
Definition VPlanCFG.h:216
iterator_range< df_iterator< VPBlockDeepTraversalWrapper< VPBlockBase * > > > vp_depth_first_deep(VPBlockBase *G)
Returns an iterator range to traverse the graph starting at G in depth-first order while traversing t...
Definition VPlanCFG.h:243
detail::concat_range< ValueT, RangeTs... > concat(RangeTs &&...Ranges)
Returns a concatenated range across two or more ranges.
Definition STLExtras.h:1152
uint64_t PowerOf2Ceil(uint64_t A)
Returns the power of two which is greater than or equal to the given value.
Definition MathExtras.h:396
auto dyn_cast_or_null(const Y &Val)
Definition Casting.h:759
bool any_of(R &&range, UnaryPredicate P)
Provide wrappers to std::any_of which take ranges instead of having to pass begin/end explicitly.
Definition STLExtras.h:1712
auto reverse(ContainerTy &&C)
Definition STLExtras.h:408
iterator_range< po_iterator< VPBlockDeepTraversalWrapper< VPBlockBase * > > > vp_post_order_deep(VPBlockBase *G)
Returns an iterator range to traverse the graph starting at G in post order while traversing through ...
Definition VPlanCFG.h:236
void sort(IteratorTy Start, IteratorTy End)
Definition STLExtras.h:1624
bool none_of(R &&Range, UnaryPredicate P)
Provide wrappers to std::none_of which take ranges instead of having to pass begin/end explicitly.
Definition STLExtras.h:1719
SmallVector< ValueTypeFromRangeType< R >, Size > to_vector(R &&Range)
Given a range of type R, iterate the entire range and return a SmallVector with elements of the vecto...
bool isa(const From &Val)
isa<X> - Return true if the parameter to the template is an instance of one of the template type argu...
Definition Casting.h:548
auto drop_end(T &&RangeOrContainer, size_t N=1)
Return a range covering RangeOrContainer with the last N elements excluded.
Definition STLExtras.h:325
RecurKind
These are the kinds of recurrences that we support.
@ Mul
Product of integers.
@ Sub
Subtraction of integers.
@ Add
Sum of integers.
@ AddChainWithSubs
A chain of adds and subs.
FunctionAddr VTableAddr Next
Definition InstrProf.h:141
auto count(R &&Range, const E &Element)
Wrapper function around std::count to count the number of times an element Element occurs in the give...
Definition STLExtras.h:1934
DWARFExpression::Operation Op
auto count_if(R &&Range, UnaryPredicate P)
Wrapper function around std::count_if to count the number of times an element satisfying a given pred...
Definition STLExtras.h:1941
decltype(auto) cast(const From &Val)
cast<X> - Return the argument parameter cast to the specified type.
Definition Casting.h:565
LLVM_ABI BasicBlock * SplitBlock(BasicBlock *Old, BasicBlock::iterator SplitPt, DominatorTree *DT, LoopInfo *LI=nullptr, MemorySSAUpdater *MSSAU=nullptr, const Twine &BBName="", bool Before=false)
Split the specified block at the specified instruction.
auto find_if(R &&Range, UnaryPredicate P)
Provide wrappers to std::find_if which take ranges instead of having to pass begin/end explicitly.
Definition STLExtras.h:1738
Type * getLoadStoreType(const Value *I)
A helper function that returns the type of a load or store instruction.
@ DataAndControlFlowWithoutRuntimeCheck
Use predicate to control both data and control flow, but modify the trip count so that a runtime over...
hash_code hash_combine(const Ts &...args)
Combine values into a single hash_code.
Definition Hashing.h:592
bool equal(L &&LRange, R &&RRange)
Wrapper function around std::equal to detect if pair-wise elements between two ranges are the same.
Definition STLExtras.h:2068
Type * toVectorTy(Type *Scalar, ElementCount EC)
A helper function for converting Scalar types to vector types.
@ Default
The result values are uniform if and only if all operands are uniform.
Definition Uniformity.h:20
std::unique_ptr< VPlan > VPlanPtr
Definition VPlan.h:77
constexpr detail::IsaCheckPredicate< Types... > IsaPred
Function object wrapper for the llvm::isa type check.
Definition Casting.h:836
hash_code hash_combine_range(InputIteratorT first, InputIteratorT last)
Compute a hash_code for a sequence of values.
Definition Hashing.h:466
#define N
This struct is a compact representation of a valid (non-zero power of two) alignment.
Definition Alignment.h:39
An information struct used to provide DenseMap with the various necessary components for a given valu...
Incoming for lane maks phi as machine instruction, incoming register Reg and incoming block Block are...
A range of powers-of-2 vectorization factors with fixed start and adjustable end.
Struct to hold various analysis needed for cost computations.
A recipe for handling first-order recurrence phis.
Definition VPlan.h:2287
A recipe for widening load operations with vector-predication intrinsics, using the address to load f...
Definition VPlan.h:3246
A recipe for widening load operations, using the address to load from and an optional mask.
Definition VPlan.h:3206
A recipe for widening select instructions.
Definition VPlan.h:1719
A recipe for widening store operations with vector-predication intrinsics, using the value to store,...
Definition VPlan.h:3328
A recipe for widening store operations, using the stored value, the address to store to and an option...
Definition VPlan.h:3286
static void materializeBroadcasts(VPlan &Plan)
Add explicit broadcasts for live-ins and VPValues defined in Plan's entry block if they are used as v...
static void materializeBackedgeTakenCount(VPlan &Plan, VPBasicBlock *VectorPH)
Materialize the backedge-taken count to be computed explicitly using VPInstructions.
static void optimizeInductionExitUsers(VPlan &Plan, DenseMap< VPValue *, VPValue * > &EndValues, ScalarEvolution &SE)
If there's a single exit block, optimize its phi recipes that use exiting IV values by feeding them p...
static void canonicalizeEVLLoops(VPlan &Plan)
Transform EVL loops to use variable-length stepping after region dissolution.
static void dropPoisonGeneratingRecipes(VPlan &Plan, const std::function< bool(BasicBlock *)> &BlockNeedsPredication)
Drop poison flags from recipes that may generate a poison value that is used after vectorization,...
static void createAndOptimizeReplicateRegions(VPlan &Plan)
Wrap predicated VPReplicateRecipes with a mask operand in an if-then region block and remove the mask...
static void createInterleaveGroups(VPlan &Plan, const SmallPtrSetImpl< const InterleaveGroup< Instruction > * > &InterleaveGroups, VPRecipeBuilder &RecipeBuilder, const bool &ScalarEpilogueAllowed)
static bool runPass(bool(*Transform)(VPlan &, ArgsTy...), VPlan &Plan, typename std::remove_reference< ArgsTy >::type &...Args)
Helper to run a VPlan transform Transform on VPlan, forwarding extra arguments to the transform.
static void addBranchWeightToMiddleTerminator(VPlan &Plan, ElementCount VF, std::optional< unsigned > VScaleForTuning)
Add branch weight metadata, if the Plan's middle block is terminated by a BranchOnCond recipe.
static void materializeBuildVectors(VPlan &Plan)
Add explicit Build[Struct]Vector recipes that combine multiple scalar values into single vectors.
static DenseMap< const SCEV *, Value * > expandSCEVs(VPlan &Plan, ScalarEvolution &SE)
Expand VPExpandSCEVRecipes in Plan's entry block.
static void convertToConcreteRecipes(VPlan &Plan)
Lower abstract recipes to concrete ones, that can be codegen'd.
static void convertToAbstractRecipes(VPlan &Plan, VPCostContext &Ctx, VFRange &Range)
This function converts initial recipes to the abstract recipes and clamps Range based on cost model f...
static void materializeConstantVectorTripCount(VPlan &Plan, ElementCount BestVF, unsigned BestUF, PredicatedScalarEvolution &PSE)
static void addExplicitVectorLength(VPlan &Plan, const std::optional< unsigned > &MaxEVLSafeElements)
Add a VPEVLBasedIVPHIRecipe and related recipes to Plan and replaces all uses except the canonical IV...
static void replaceSymbolicStrides(VPlan &Plan, PredicatedScalarEvolution &PSE, const DenseMap< Value *, const SCEV * > &StridesMap)
Replace symbolic strides from StridesMap in Plan with constants when possible.
static void removeBranchOnConst(VPlan &Plan)
Remove BranchOnCond recipes with true or false conditions together with removing dead edges to their ...
static void removeDeadRecipes(VPlan &Plan)
Remove dead recipes from Plan.
static void materializeVectorTripCount(VPlan &Plan, VPBasicBlock *VectorPHVPBB, bool TailByMasking, bool RequiresScalarEpilogue)
Materialize vector trip count computations to a set of VPInstructions.
static void simplifyRecipes(VPlan &Plan)
Perform instcombine-like simplifications on recipes in Plan.
static LLVM_ABI_FOR_TEST bool tryToConvertVPInstructionsToVPRecipes(VPlanPtr &Plan, function_ref< const InductionDescriptor *(PHINode *)> GetIntOrFpInductionDescriptor, const TargetLibraryInfo &TLI)
Replaces the VPInstructions in Plan with corresponding widen recipes.
static void handleUncountableEarlyExit(VPBasicBlock *EarlyExitingVPBB, VPBasicBlock *EarlyExitVPBB, VPlan &Plan, VPBasicBlock *HeaderVPBB, VPBasicBlock *LatchVPBB)
Update Plan to account for the uncountable early exit from EarlyExitingVPBB to EarlyExitVPBB by.
static void clearReductionWrapFlags(VPlan &Plan)
Clear NSW/NUW flags from reduction instructions if necessary.
static void cse(VPlan &Plan)
Perform common-subexpression-elimination on Plan.
static void addActiveLaneMask(VPlan &Plan, bool UseActiveLaneMaskForControlFlow, bool DataAndControlFlowWithoutRuntimeCheck)
Replace (ICMP_ULE, wide canonical IV, backedge-taken-count) checks with an (active-lane-mask recipe,...
static void optimize(VPlan &Plan)
Apply VPlan-to-VPlan optimizations to Plan, including induction recipe optimizations,...
static void dissolveLoopRegions(VPlan &Plan)
Replace loop regions with explicit CFG.
static void narrowInterleaveGroups(VPlan &Plan, ElementCount VF, unsigned VectorRegWidth)
Try to convert a plan with interleave groups with VF elements to a plan with the interleave groups re...
static void truncateToMinimalBitwidths(VPlan &Plan, const MapVector< Instruction *, uint64_t > &MinBWs)
Insert truncates and extends for any truncated recipe.
static bool adjustFixedOrderRecurrences(VPlan &Plan, VPBuilder &Builder)
Try to have all users of fixed-order recurrences appear after the recipe defining their previous valu...
static void optimizeForVFAndUF(VPlan &Plan, ElementCount BestVF, unsigned BestUF, PredicatedScalarEvolution &PSE)
Optimize Plan based on BestVF and BestUF.
static void materializeVFAndVFxUF(VPlan &Plan, VPBasicBlock *VectorPH, ElementCount VF)
Materialize VF and VFxUF to be computed explicitly using VPInstructions.