parallel.cc 25.3 KB
Newer Older
1
2
3
4
5
6
7
/*!
 * \file op/parallel.cc
 * \brief Define Parallel for operator
 */

#include "parallel.h"

8
#include <algorithm>
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
#include <tvm/tir/op.h>

#include "../layout/utils.h"
#include "../target/utils.h"
#include "../transform/loop_partition.h"
#include "../transform/loop_vectorize.h"

namespace tvm {
namespace tl {

using namespace tir;

namespace attr {
/*! \brief Mark that how the loop is vectorized. */
constexpr const char *coalesced_width = "coalesced_width";
24
} // namespace attr
25

26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
// ProveFragmentContains checks whether the threads that access elements of a
// smaller fragment (small_frag) are a subset of the threads that access
// elements of a larger fragment (large_frag) for any given loop index. This
// function ensures that if the small fragment's layout corresponds to the loop
// itself, accessing the large fragment's elements is valid. Additionally, if
// small is updated to large, the originally valid access remains valid. The
// proof is performed by:
//
// 1. Defining a variable `rep_small` to represent the replicate index of the
//    small fragment that is being checked.
// 2. Using the `small_frag_indices` and `rep_small` to derive the thread
// accessing
//    the element in the small fragment.
// 3. Using `large_frag_indices` to derive the physical index of the large
// fragment
//    along with the thread information, and then feeding these into the inverse
//    of the large fragment to obtain the logical index and replicate index.
// 4. Verifying the mapping by checking whether the computed thread using the
// inverse
//    layout corresponds to the original thread calculated for the small
//    fragment. If they don't match, this indicates that the inverse layout's
//    domain does not include the thread and thus the access is invalid.
bool ProveFragmentContains(Fragment small_frag, Fragment large_frag,
                           Array<PrimExpr> small_frag_indices,
                           Array<PrimExpr> large_frag_indices,
                           arith::Analyzer &analyzer_) {
  Var rep_small("__checking_frag_contains_rep");
  analyzer_.Bind(rep_small,
                 Range(IntImm(small_frag->ReplicateExtent()->dtype, 0),
                       small_frag->ReplicateExtent()),
                 true); // Bind the replicate extent of small_frag.
  // Derive thread for small_frag.
  auto thread = small_frag->ForwardThread(small_frag_indices, rep_small);

  // Get physical index and thread for large_frag.
  auto large_frag_physical_and_thread = large_frag->Forward(large_frag_indices);
  // Add small_frag's thread to the large fragment's thread info.
  large_frag_physical_and_thread.push_back(thread);
  // Get the inverse of the large fragment.
  auto inv_large_frag = large_frag->Inverse();
  // Compute logical index and replicate index using inverse layout.
  auto inv_large_frag_logical_and_rep =
      inv_large_frag->Forward(large_frag_physical_and_thread);

  // Extract replicate index from the result.
  auto inv_large_frag_rep =
      inv_large_frag_logical_and_rep[inv_large_frag_logical_and_rep.size() - 1];

  // Calculate thread based on the logical index and replicate index.
  auto check_thread =
      large_frag->ForwardThread(large_frag_indices, inv_large_frag_rep);

  // Simplify the difference between the threads.
  auto diff = analyzer_.Simplify(thread - check_thread);
  // If the difference is zero, the threads match and the access is valid.
  return is_zero(diff);
}

84
class IfBufferRemapLoopGenerator : public StmtExprMutator {
85
public:
86
87
88
89
90
91
  static For run(Stmt stmt, Map<Buffer, Buffer> buffer_remap,
                 Map<Buffer, Layout> layout_map) {
    IfBufferRemapLoopGenerator generator(buffer_remap, layout_map);
    return Downcast<For>(generator(std::move(stmt)));
  }

92
93
94
private:
  IfBufferRemapLoopGenerator(Map<Buffer, Buffer> buffer_remap,
                             Map<Buffer, Layout> layout_map)
95
96
      : buffer_remap_(buffer_remap), layout_map_(layout_map) {}

97
  PrimExpr VisitExpr_(const BufferLoadNode *op) final {
98
99
100
101
102
103
104
105
106
107
108
    auto load = Downcast<BufferLoad>(StmtExprMutator::VisitExpr_(op));

    if (buffer_remap_.count(load->buffer)) {
      auto new_indices = layout_map_[load->buffer]->Forward(load->indices);
      auto new_buffer = buffer_remap_[load->buffer];

      return BufferLoad(new_buffer, new_indices);
    }
    return load;
  }

109
  Stmt VisitStmt_(const BufferStoreNode *op) final {
110
111
112
113
114
115
116
117
118
119
120
121
122
    auto store = Downcast<BufferStore>(StmtExprMutator::VisitStmt_(op));
    if (buffer_remap_.count(store->buffer)) {
      auto new_indices = layout_map_[store->buffer]->Forward(store->indices);
      auto new_buffer = buffer_remap_[store->buffer];
      return BufferStore(new_buffer, store->value, new_indices);
    }
    return store;
  }

  Map<Buffer, Buffer> buffer_remap_;
  Map<Buffer, Layout> layout_map_;
};

123
124
125
126
127
128
129
130
/**
 * @brief Handle a parallel For node during traversal, collecting loop metadata.
 *
 * Visits a parallel loop, asserts the loop is parallel, records a data-parallel
 * IterVar for the loop, binds the loop variable range into the analyzer scope,
 * and extracts any reducer information from the loop's annotations into the
 * visitor's reducer_info_map_. Continues traversal into the loop body.
 */
131
void ParallelLoopNestVisitor::VisitStmt_(const ForNode *op) {
132
133
134
135
136
137
138
  if (op->kind == ForKind::kParallel)
    p->loop_vars_.push_back(IterVar(Range(op->min, op->extent), op->loop_var,
                                    IterVarType::kDataPar));
  else
    p->inner_vars_.Set(op->loop_var,
                       IterVar(Range(op->min, op->extent), op->loop_var,
                               IterVarType::kOrdered));
139
  p->analyzer_.Bind(op->loop_var, Range::FromMinExtent(op->min, op->extent));
140
141
142
143
144
145
  auto reducer_info_map =
      op->annotations.Get(attr::kReducerInfo)->as<Map<Var, ReducerInfo>>();
  if (reducer_info_map) {
    for (auto &&[buffer, info] : reducer_info_map.value())
      p->reducer_info_map_.Set(buffer, info);
  }
146
147
148
  StmtExprVisitor::VisitStmt_(op);
}

149
void ParallelLoopNestVisitor::VisitStmt_(const BufferStoreNode *op) {
150
151
152
  if (op->buffer.scope() == "local.fragment") {
    if (p->indice_map_.find(op->buffer) != p->indice_map_.end()) {
      ICHECK(StructuralEqual()(p->indice_map_.at(op->buffer), op->indices))
153
154
          << op->buffer << ": " << op->indices << " and "
          << p->indice_map_.at(op->buffer);
155
156
157
158
159
160
161
162
    } else {
      p->indice_map_.Set(op->buffer, op->indices);
    }
    p->buffer_is_write_.insert(op->buffer);
  }
  StmtExprVisitor::VisitStmt_(op);
}

163
void ParallelLoopNestVisitor::VisitExpr_(const BufferLoadNode *op) {
164
165
166
  if (op->buffer.scope() == "local.fragment") {
    if (p->indice_map_.find(op->buffer) != p->indice_map_.end()) {
      ICHECK(StructuralEqual()(p->indice_map_.at(op->buffer), op->indices))
167
168
          << op->buffer << ": " << op->indices << " and "
          << p->indice_map_.at(op->buffer);
169
170
171
172
173
174
175
    } else {
      p->indice_map_.Set(op->buffer, op->indices);
    }
  }
  StmtExprVisitor::VisitExpr_(op);
}

176
177
178
179
180
181
182
183
184
185
186
187
188
ParallelOpNode::ParallelOpNode(For root) : root_(root), V(this) {
  V.VisitStmt(root);
}

TileOperator ParallelOpNode::Clone() const {
  auto op = make_object<ParallelOpNode>(*this);
  return ParallelOp(op);
}

Stmt ParallelOpNode::Lower(const LowerArgs &T,
                           arith::Analyzer *analyzer) const {
  return root_;
}
189

190
bool ParallelOpNode::IsCommonAccessIndice(const Buffer &buffer) const {
191
  auto common_indice = loop_vars_.Map([](const auto &iv) { return iv->var; });
192
193
194
  return StructuralEqual()(indice_map_[buffer], common_indice);
}

195
196
/*! \brief Infer the layout for parallel operations based on different inference
 * levels
197
 *
198
199
200
201
202
 * The inference level controls how aggressively we try to infer and optimize
 * layouts:
 * - kStrict (2): Most conservative level. Only allows explicitly defined
 * layouts. Returns empty layout map if loop_layout_ is not already defined.
 *                Used when exact layout control is required.
203
 *
204
205
206
 * - kCommon (1): Intermediate level between strict and free.
 *                Allows common layout patterns while maintaining some
 * constraints.
207
 *
208
209
210
211
 * - kFree (0):   Most permissive level. Allows maximum optimization freedom.
 *                Will attempt layout inference even without source buffers.
 *                Can generate new layouts based on vectorization and thread
 * bounds. Used when maximum performance optimization is desired.
212
 */
213
214
LayoutMap ParallelOpNode::InferLayout(const LayoutInferArgs &T,
                                      InferLevel level) const {
215
216
  if (loop_layout_.defined())
    return {};
217
218
  if (level == InferLevel::kStrict) {
    LayoutMap results;
219
    // Deduce buffers that should be complicated replicated.
220
    // For example:
221
    // for i in T.Parallel(m):
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
    //   fragment[0] = x[i]
    // then fragment[0] must be replicated on all threads.
    for (const auto &[buffer, indices] : indice_map_) {
      if (T.layout_map.count(buffer)) {
        continue;
      }
      if (buffer.scope() != "local.fragment")
        continue;

      // Check if all indices are zero
      bool all_indices_zero = true;
      for (const auto &index : indices) {
        if (const auto *imm = index.as<IntImmNode>()) {
          if (imm->value != 0) {
            all_indices_zero = false;
            LOG(FATAL)
                << "Fragment buffer access with non-zero index [" << imm->value
                << "] is not supported. "
                << "Only fragment[0] access is allowed within T.Parallel loop.";
          }
        } else {
          // Non-constant index, not all zero
          all_indices_zero = false;
        }
      }

      // Only set layout if all indices are zero
      if (all_indices_zero) {
        Array<IterVar> forward_vars;
        for (const auto &s : buffer->shape) {
          forward_vars.push_back(
              IterVar(Range(0, s), Var(), IterVarType::kDataPar));
        }
        Array<PrimExpr> forward_index;
        for (const auto &iv : forward_vars) {
          forward_index.push_back(iv->var);
        }
        Var rep;
        auto rep_iter =
            IterVar({0, T.thread_bounds->extent}, rep, IterVarType::kDataPar);

        const PrimExpr &forward_thread = rep;
        results.Set(buffer, Fragment(forward_vars, forward_index,
                                     forward_thread, rep_iter));
      }
    }
    return results;
  }
  auto buffer_is_completed_replicated = [&](const Buffer &buffer) {
    if (buffer.scope() != "local.fragment")
      return false;
    auto frag = T.layout_map[buffer].as<Fragment>().value();
    // buffer indices should be IntImm
    for (const auto &index : indice_map_[buffer]) {
      if (!index.as<IntImmNode>()) {
        return false;
      } else if (index.as<IntImmNode>()->value != 0) {
        LOG(FATAL) << "buffer " << buffer << " is not completed replicated";
      }
    }
    return frag->IsCompletedReplicated();
  };
  // Collect fragment buffers with const index and all fragment_buffers
  std::vector<Buffer> const_index_fragment_buffer, fragment_buffers;
  for (const auto &[buffer, indices] : indice_map_) {
    if (buffer.scope() != "local.fragment")
      continue;
    fragment_buffers.push_back(buffer);

    bool is_const_index = true;
    for (const auto &index : indices) {
      if (!index.as<IntImmNode>()) {
        is_const_index = false;
        break;
      }
    }
    if (is_const_index) {
      const_index_fragment_buffer.push_back(buffer);
    }
  }

  // Determine if common layout propagation should be applied.
  // If there are fragment buffers with non-constant indices, we need to
  // propagate the common layout pattern to ensure consistency across all
  // fragments. Example cases:
  //   - Need propagation: frag_a[0] = T.min(frag_a[0], frag_b[i])
  //     (const index frag_a interacts with non-const index frag_b)
  //   - No propagation needed: shared_a[i] = frag_a[0]
  //     (const index frag_a with non-fragment buffer)
311

312
  bool allow_layout_propgate =
313
314
      const_index_fragment_buffer.empty() ||
      (fragment_buffers.size() > const_index_fragment_buffer.size());
315
316
317

  // Step 1: try to infer loop's partition from a source fragment
  Buffer source_buffer, read_source_buffer;
318
319
  Buffer replicated_write_buffer; // Backup: fully replicated write buffer

320
  for (const auto &[buffer, indices] : indice_map_) {
321
    if (T.layout_map.count(buffer)) {
322
323
324
325
326
      // skip reducers with rep=ALL
      if (auto info = reducer_info_map_.Get(buffer->data);
          info && info.value()->rep == ReducerRepType::ALL)
        continue;

327
      auto frag = T.layout_map[buffer].as<Fragment>().value();
328
329
      bool is_fully_replicated = buffer_is_completed_replicated(buffer);

330
      if (buffer_is_write_.count(buffer)) {
331
        source_buffer = buffer;
332
333
334
335
      } else {
        // Keep the buffer with largest number of indices
        // (which means the inference based on that buffer is more accurate)
        // as read_source_buffer to get more accurate layout
336
337
338
339
340
        // if the buffer is completed replicated, we don't need to infer the
        // layout from this buffer.
        if ((!read_source_buffer.defined() ||
             indice_map_[buffer].size() >
                 indice_map_[read_source_buffer].size())) {
341
342
          read_source_buffer = buffer;
        }
343
344
345
346
        // If the buffer is not replicated and shape is equal to the
        // source_buffer, use it as source_buffer because the layout inference
        // is more accurate
        if (is_one(frag->ReplicateExtent()) && !source_buffer.defined()) {
347
348
          source_buffer = buffer;
        }
349
      }
350
351
    }
  }
352
  auto compute_loop_layout_from_buffer = [&](const Buffer &buffer) {
353
    Fragment src_layout = T.layout_map[buffer].as<Fragment>().value();
354
355
    DLOG(INFO) << "[compute_loop_layout_from_buffer] infer from buffer `"
               << buffer << "` of layout " << src_layout->DebugOutput() << '\n';
356

357
    Fragment result;
358
    if (IsCommonAccessIndice(buffer)) {
359
      result = src_layout;
360
361
    } else {
      Var rep;
362
363
364
365
      auto rep_iter = IterVar({0, src_layout->ReplicateExtent()}, rep,
                              IterVarType::kDataPar);
      PrimExpr loop_var_to_thread =
          src_layout->ForwardThread(indice_map_[buffer], rep);
366
      loop_var_to_thread = analyzer_.Simplify(loop_var_to_thread);
367
368
369
370
371
372
373
374
375
      PostOrderVisit(loop_var_to_thread, [&](const ObjectRef &objref) {
        if (auto opt_var = objref.as<Var>();
            opt_var && inner_vars_.count(*opt_var)) {
          std::ostringstream oss;
          oss << "loop_var_to_thread = " << loop_var_to_thread
              << "contains inner var" << *opt_var;
          throw LayoutConflictException(oss.str());
        }
      });
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390

      try {
        result = Fragment(loop_vars_, {}, loop_var_to_thread, rep_iter)
                     ->BindThreadRange(T.thread_bounds);
      } catch (const tvm::runtime::Error &err) {
        std::ostringstream msg;
        msg << "Layout inference for buffer `" << buffer->name
            << "` failed inside `T.parallel` loop.";

        msg << "\nUnderlying TVM error: " << err.what();
        msg << "\nProblematic loop AST:\n " << root_;
        msg << "\nHint: ensure the loop extent divides the thread binding or "
               "adjust the fragment mapping.";
        LOG(FATAL) << msg.str();
      }
391
    }
392
393
394
    DLOG(INFO) << "[compute_loop_layout_from_buffer] ... and get "
               << result->DebugOutput() << '\n';
    return result;
395
  };
396
397
398
399
400
401
402
403

  // Try to infer loop layout from buffers in order of preference:
  // 1. Non-replicated write buffer (most reliable)
  // 2. Non-replicated read buffer
  // 3. Fully replicated write buffer (backup, may cause issues)
  // 4. Free inference mode (no source buffer)

  if (source_buffer.defined() && allow_layout_propgate) {
404
405
    loop_layout_ = compute_loop_layout_from_buffer(source_buffer);
  } else if (level == InferLevel::kFree) {
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
    // For free layout inference
    // If replication exists and buffer has cross-thread shared memory access,
    // add predicate
    bool has_cross_thread_access = false;
    PostOrderVisit(root_, [&](const ObjectRef &obj) {
      if (const auto *store = obj.as<BufferStoreNode>()) {
        // check if scope is shared or global
        if (store->buffer.scope() == "shared" ||
            store->buffer.scope() == "shared.dyn" ||
            store->buffer.scope() == "global") {
          has_cross_thread_access = true;
        }
      } else if (const auto *load = obj.as<BufferLoadNode>()) {
        // check if scope is shared or global
        if (load->buffer.scope() == "shared" ||
            load->buffer.scope() == "shared.dyn" ||
            load->buffer.scope() == "global") {
          has_cross_thread_access = true;
        }
      }
    });

    // check if loop body contains a "pure" buffer store (i.e., direct
    // assignment, not compound update)
430
431
432
433
    std::vector<Buffer> store_shared_global_buffers, store_fragment_buffers;
    // Buffers that scope is above fragments.
    // global, shared, shared.dyn
    // which can be used to analysis replicate case
434
435
    PostOrderVisit(root_, [&](const ObjectRef &obj) {
      if (const auto *store = obj.as<BufferStoreNode>()) {
436
437
438
439
440
441
        auto buffer = store->buffer;
        if (buffer.scope() == "shared" || buffer.scope() == "shared.dyn" ||
            buffer.scope() == "global") {
          store_shared_global_buffers.emplace_back(buffer);
        } else if (buffer.scope() == "local.fragment") {
          store_fragment_buffers.emplace_back(buffer);
442
443
444
        }
      }
    });
445
    if (read_source_buffer.defined() && allow_layout_propgate) {
446
      loop_layout_ = compute_loop_layout_from_buffer(read_source_buffer);
447
448
449
450
    }

    if (!loop_layout_.defined()) {
      // No source buffer available, use free mode inference
451
452
      // Vectorize Size must be aware of the buffer_remap
      // As the pass will do post processing to the layout
453
454
      auto maybe_remapped_root_ =
          IfBufferRemapLoopGenerator::run(root_, T.buffer_remap, T.layout_map);
455
456
      int vector_size = GetVectorizeSize(maybe_remapped_root_);

457
458
      DLOG(INFO) << "[PlanLoopPartition] vector_size = " << vector_size << '\n';

459
460
461
462
      PrimExpr loop_total_size = 1;
      for (Stmt l = root_; l.as<For>().has_value();
           l = l.as<For>().value()->body)
        loop_total_size = loop_total_size * l.as<For>().value()->extent;
463
464
      DLOG(INFO) << "[PlanLoopPartition] loop_total_size = " << loop_total_size
                 << '\n';
465
466
467
468
469
      while (!analyzer_.CanProve(
                 floormod(loop_total_size,
                          T.thread_bounds->extent * vector_size) == 0) &&
             vector_size > 1)
        vector_size /= 2;
470
471
      DLOG(INFO) << "[PlanLoopPartition] after adjust: vector_size = "
                 << vector_size << '\n';
472

473
      // Check if coalesced_width is defined
474
475
      if (auto coalesced_width =
              root_->annotations.Get(tl::attr::coalesced_width)) {
476
        if (const auto *imm = coalesced_width->as<IntImmNode>()) {
477
478
479
          int expected = imm->value;
          // Verify that vector_size is divisible by expected
          if (vector_size % expected != 0) {
480
481
            LOG(FATAL) << "Vector size " << vector_size
                       << " is not divisible by coalesced width " << expected;
482
483
484
485
486
487
          }
          vector_size = expected;
        } else {
          LOG(FATAL) << "coalesced_width should be an IntImmNode.";
        }
      }
488
489
490
      DLOG(INFO) << "[PlanLoopPartition] root_ = " << root_
                 << " ############# vector_size = " << vector_size
                 << ", thread_bounds = " << T.thread_bounds << '\n';
491
      loop_layout_ = PlanLoopPartition(root_, vector_size, T.thread_bounds);
492
493
      DLOG(INFO) << "[PlanLoopPartition] loop_layout_ = "
                 << loop_layout_->DebugOutput() << '\n';
494
    }
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561

    // Lambda that guards replicated accesses:
    // - When a loop layout replicates a fragment buffer (rep > 1), each thread
    //   observes the same fragment elements. Blindly storing to shared/global
    //   memory in that case would add the same value multiple times.
    // - We therefore restrict the store so that only the replica with rep == 0
    //   performs the update (e.g. global[i] += fragment[i] only fires once).
    // Trigger conditions for this guard:
    // 1) There are cross-thread stores targeting shared/global memory (no
    //    fragment stores in this branch; atomic_add and similar remain TODO).
    // 2) The loop layout replicate extent is greater than 1, inferred from the
    //    thread bounds captured in the layout.

    [this, &store_shared_global_buffers, &store_fragment_buffers,
     &has_cross_thread_access, &const_index_fragment_buffer, &T]() {
      if (is_one(loop_layout_->ReplicateExtent()))
        return;
      if (!has_cross_thread_access)
        return;

      if (!store_fragment_buffers.empty()) {
        // Iterate replicated fragment stores: when the fragment index is a
        // constant (e.g. fragment[0]), every thread touches the same slot, so
        // the rep == 0 predicate is unnecessary. Example: for i in
        // T.Parallel(...):
        //   shared[i] = ...
        //   fragment[0] = ...
        bool replicate_is_from_dynamic_index_fragment = false;
        for (const auto &fragment : store_fragment_buffers) {
          if (!T.layout_map.count(fragment)) {
            continue;
          }

          auto fragment_layout = T.layout_map[fragment].as<Fragment>().value();
          if (is_one(fragment_layout->ReplicateExtent()))
            continue;

          if (analyzer_.CanProveEqual(fragment_layout->ReplicateExtent(),
                                      loop_layout_->ReplicateExtent()))
            continue;
          if (std::find(const_index_fragment_buffer.begin(),
                        const_index_fragment_buffer.end(),
                        fragment) == const_index_fragment_buffer.end()) {
            replicate_is_from_dynamic_index_fragment = true;
          }
        }

        if (!replicate_is_from_dynamic_index_fragment)
          return;

        ICHECK(store_shared_global_buffers.empty())
            << "Invalid layout: cannot have both fragment and shared store "
               "buffers "
               "in replicated loop layout.";
        return;
      } else {
        // Now, store is global or shared
        // or T.call_extern or T.call_intrin ...
        auto inv = loop_layout_->Inverse();
        Array<PrimExpr> fwd;
        for (size_t i = 0; i < loop_layout_->OutputDim(); i++)
          fwd.push_back(0);
        fwd.push_back(InputPlaceholder(0) - T.thread_bounds->min);
        auto rep = inv->Forward(fwd).back();
        AddPredicate(EQ(rep, 0));
      }
    }();
562
563
564
  } else {
    return {};
  }
565

566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
  PrimExpr loop_thread_extent = loop_layout_->ThreadExtent();

  auto block_size = T.thread_bounds->extent;
  if (loop_layout_.defined()) {
    if (loop_layout_->ThreadRange().defined()) {
      auto thread_range = loop_layout_->ThreadRange();
      block_size = thread_range->extent;
      AddPredicate(GE(InputPlaceholder(0), thread_range->min));
      AddPredicate(
          LT(InputPlaceholder(0), thread_range->min + thread_range->extent));
    }
  }

  if (!analyzer_.CanProveEqual(loop_thread_extent, block_size)) {
    AddPredicate(
        LT(InputPlaceholder(0), loop_thread_extent + T.thread_bounds->min));
  }

584
  // Step 2: Check that the loop's partition can correctly align with all source
585
586
  // fragment, and infer layout only when it's not yet layout-ed
  LayoutMap results;
587
  for (const auto &[buffer, _] : indice_map_) {
588
589
    if (T.layout_map.count(buffer)) {
      auto fragment = T.layout_map[buffer].as<Fragment>().value();
590
591
      auto vars =
          loop_vars_.Map([](const IterVar &iv) { return PrimExpr(iv->var); });
592
593
594
595
      if (!ProveFragmentContains(loop_layout_, fragment, vars,
                                 indice_map_[buffer], analyzer_)) {
        std::ostringstream oss;
        oss << "Layout infer conflict between " << buffer << " and "
596
597
598
            << source_buffer << " in T.Parallel loop:" << '\n'
            << "    loop " << loop_layout_->DebugOutput() << '\n'
            << "    fragment " << fragment->DebugOutput() << '\n';
599
        throw LayoutConflictException(oss.str());
600
      }
601
602
603
604
    } else {
      auto dst_layout =
          CompleteBufferFragment(buffer)->BindThreadRange(T.thread_bounds);
      results.Set(buffer, dst_layout);
605
    }
606
607
608
609
  }
  return results;
}

610
Optional<PrimExpr> ParallelOpNode::GetPredicate(Var thread_var) const {
611
612
613
  if (predicate_.defined()) {
    return Substitute(predicate_.value(), {{InputPlaceholder(0), thread_var}});
  } else {
614
    return std::nullopt;
615
616
617
  }
}

618
Fragment ParallelOpNode::CompleteBufferFragment(const Buffer &buffer) const {
619
  ICHECK(loop_layout_.defined());
620
  if (IsCommonAccessIndice(buffer)) {
621
    return loop_layout_;
622
  }
623
624
  PrimExpr rep_b = MakeFlattenedExpression(
      DivideUnusedIterators(indice_map_[buffer], loop_vars_, &analyzer_));
625
626
627
  auto bijective_indice = indice_map_[buffer];
  bijective_indice.push_back(rep_b);
  Layout ind_inv = Layout(loop_vars_, bijective_indice)->Inverse();
628
629
  PrimExpr indice_rep_extent =
      ind_inv->InputShape().back(); // this is the size of rep_b
630
631
632
633
634
635
636
637
  PrimExpr loop_rep_extent = loop_layout_->ReplicateExtent();
  PrimExpr dest_buffer_rep_extent = indice_rep_extent * loop_rep_extent;
  Array<PrimExpr> fwd;
  for (size_t i = 0; i < buffer->shape.size(); i++) {
    fwd.push_back(InputPlaceholder(i));
  }
  fwd.push_back(FloorMod(ReplicationPlaceholder(), indice_rep_extent));
  PrimExpr thd_b = loop_layout_->ForwardThread(
638
639
      ind_inv->Forward(fwd),
      FloorDiv(ReplicationPlaceholder(), indice_rep_extent));
640
641
  return Fragment(buffer->shape, {}, thd_b, dest_buffer_rep_extent,
                  std::nullopt)
642
643
644
      ->CondenseReplicateVar();
}

645
646
TVM_FFI_STATIC_INIT_BLOCK({ ParallelOpNode::RegisterReflection(); });

647
648
} // namespace tl
} // namespace tvm