graph_helper.cc 16.8 KB
Newer Older
X
better  
Xin Pan 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14
/* Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.

Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

    http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License. */

C
chengduo 已提交
15
#include "paddle/fluid/framework/ir/graph_helper.h"
16
#include <queue>
Y
Yan Chunwei 已提交
17
#include <stack>
18
#include "paddle/fluid/framework/op_proto_maker.h"
X
better  
Xin Pan 已提交
19

20
DECLARE_bool(convert_all_blocks);
21 22 23
PADDLE_DEFINE_EXPORTED_string(print_sub_graph_dir, "",
                              "FLAGS_print_sub_graph_dir is used "
                              "to print the nodes of sub_graphs.");
C
chengduo 已提交
24

X
better  
Xin Pan 已提交
25 26 27 28
namespace paddle {
namespace framework {
namespace ir {
namespace {
29 30 31 32

template <class NodeComparator = ir::NodeComp>
void SortHelper(const std::map<ir::Node *, std::set<ir::Node *, NodeComparator>,
                               NodeComparator> &adj_list,
33 34
                ir::Node *node, std::unordered_set<ir::Node *> *visited,
                std::vector<ir::Node *> *ret) {
X
better  
Xin Pan 已提交
35 36 37 38
  visited->insert(node);

  for (auto adj : adj_list.at(node)) {
    if (visited->find(adj) == visited->end()) {
39
      SortHelper<NodeComparator>(adj_list, adj, visited, ret);
X
better  
Xin Pan 已提交
40 41 42
    }
  }

Y
Yan Chunwei 已提交
43
  VLOG(5) << "topology sort insert: " << node->Name() << " "
M
minqiyang 已提交
44
          << reinterpret_cast<void *>(node) << " input " << node->inputs.size();
X
better  
Xin Pan 已提交
45 46 47
  ret->push_back(node);
}

48
template <class NodeComparator = ir::NodeComp>
X
better  
Xin Pan 已提交
49 50
bool HasCircleHelper(
    ir::Node *node,
51 52
    const std::map<ir::Node *, std::set<ir::Node *, NodeComparator>,
                   NodeComparator> &adj_list,
X
better  
Xin Pan 已提交
53
    std::unordered_set<ir::Node *> *visited,
D
dzhwinter 已提交
54 55
    std::unordered_set<ir::Node *> *in_trace,
    std::vector<std::vector<ir::Node *>> *circles) {
X
better  
Xin Pan 已提交
56 57 58 59 60 61
  if (visited->find(node) == visited->end()) {
    visited->insert(node);
    in_trace->insert(node);

    for (ir::Node *in : adj_list.at(node)) {
      if (visited->find(in) == visited->end() &&
62 63
          HasCircleHelper<NodeComparator>(in, adj_list, visited, in_trace,
                                          circles)) {
X
better  
Xin Pan 已提交
64 65
        return true;
      } else if (in_trace->find(in) != in_trace->end()) {
D
dzhwinter 已提交
66 67 68 69 70 71 72 73 74 75 76 77
        if (circles != nullptr) {
          std::vector<ir::Node *> circle;
          circle.emplace_back(in);
          ir::Node *p = in;
          for (auto &adj : adj_list.at(p)) {
            if (in_trace->count(adj)) {
              circle.emplace_back(adj);
              p = adj;
            }
          }
          circles->emplace_back(circle);
        }
X
better  
Xin Pan 已提交
78 79 80 81 82 83 84 85
        return true;
      }
    }
  }
  in_trace->erase(node);
  return false;
}

86
template <class NodeComparator = ir::NodeComp>
X
Xin Pan 已提交
87
bool HasCircleInternal(
88 89
    const std::map<ir::Node *, std::set<ir::Node *, NodeComparator>,
                   NodeComparator> &adj_list,
D
dzhwinter 已提交
90
    std::vector<std::vector<ir::Node *>> *circles) {
X
better  
Xin Pan 已提交
91 92 93
  std::unordered_set<ir::Node *> visited;
  std::unordered_set<ir::Node *> in_trace;
  for (auto &adj : adj_list) {
94 95
    if (HasCircleHelper<NodeComparator>(adj.first, adj_list, &visited,
                                        &in_trace, circles)) {
X
better  
Xin Pan 已提交
96 97 98 99 100
      return true;
    }
  }
  return false;
}
X
Xin Pan 已提交
101 102 103
}  // namespace

bool HasCircle(const Graph &graph) {
D
dzhwinter 已提交
104 105 106
  return HasCircleInternal(BuildOperationAdjList(graph), nullptr);
}

107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131
bool VarDescIsConsistency(const Graph &graph) {
  std::unordered_map<std::string, std::unordered_set<ir::Node *>>
      var_name2node_set;
  for (ir::Node *node : graph.Nodes()) {
    if (node->IsVar() && node->Var()) {
      var_name2node_set[node->Var()->Name()].emplace(node);
    }
  }
  for (auto &iter : var_name2node_set) {
    auto &first_node = *iter.second.begin();
    bool is_persistable = std::any_of(iter.second.begin(), iter.second.end(),
                                      [&first_node](const ir::Node *node) {
                                        return node->Var()->Persistable();
                                      });
    if (is_persistable) {
      bool is_consistency =
          std::all_of(iter.second.begin(), iter.second.end(),
                      [&first_node](const ir::Node *node) {
                        return *node->Var() == *first_node->Var();
                      });
      if (!is_consistency) return false;
    }
  }
  return true;
}
D
dzhwinter 已提交
132 133 134
bool FindCircleSubGraph(const Graph &graph,
                        std::vector<std::vector<ir::Node *>> *circles) {
  return HasCircleInternal(BuildOperationAdjList(graph), circles);
X
Xin Pan 已提交
135
}
X
better  
Xin Pan 已提交
136

X
Xin Pan 已提交
137
std::vector<ir::Node *> TopologySortOperations(const Graph &graph) {
138 139
  std::map<ir::Node *, std::set<ir::Node *, ir::NodeComp>, ir::NodeComp>
      adj_list = BuildOperationAdjList(graph);
140 141 142
  PADDLE_ENFORCE_EQ(HasCircleInternal(adj_list, nullptr), false,
                    platform::errors::InvalidArgument(
                        "Generated graph shouldn't contain cycle."));
X
better  
Xin Pan 已提交
143 144 145 146
  std::unordered_set<ir::Node *> visited;
  std::vector<ir::Node *> ret;
  for (auto adj : adj_list) {
    if (visited.find(adj.first) == visited.end()) {
147
      SortHelper<ir::NodeComp>(adj_list, adj.first, &visited, &ret);
X
better  
Xin Pan 已提交
148 149
    }
  }
150

X
better  
Xin Pan 已提交
151 152 153
  return ret;
}

Z
Zeng Jinle 已提交
154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179
bool IsTopologySortOperationsUnique(const Graph &graph) {
  auto nodes = TopologySortOperations(graph);
  size_t n = nodes.size();
  for (size_t i = 1; i < n; ++i) {
    auto *prev_op = nodes[i - 1];
    auto *cur_op = nodes[i];

    std::unordered_set<Node *> prev_op_outputs;
    for (auto *output : prev_op->outputs) {
      prev_op_outputs.insert(output);
    }

    bool found = false;
    for (auto *input : cur_op->inputs) {
      if (prev_op_outputs.count(input) > 0) {
        found = true;
        break;
      }
    }
    if (!found) {
      return false;
    }
  }
  return true;
}

Y
Yan Chunwei 已提交
180 181 182 183 184 185 186 187 188 189 190 191
// Build operator outlink edge table.
std::map<ir::Node *, std::unordered_set<ir::Node *>> BuildOperationOutAdjList(
    const Graph &graph) {
  std::map<ir::Node *, std::unordered_set<ir::Node *>> adj_list;

  for (auto &n : graph.Nodes()) {
    if (!n->IsOp()) continue;
    if (adj_list.find(n) == adj_list.end()) {
      adj_list[n] = std::unordered_set<ir::Node *>();
    }
    for (auto &var : n->outputs) {
      for (auto &adj_n : var->outputs) {
192 193 194 195 196
        PADDLE_ENFORCE_EQ(
            adj_n->NodeType(), ir::Node::Type::kOperation,
            platform::errors::InvalidArgument(
                "Node(%s)'s type(%d) must be kOperation type.", adj_n->Name(),
                static_cast<int>(adj_n->NodeType())));
Y
Yan Chunwei 已提交
197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296
        VLOG(40) << "adj " << adj_n->Name() << reinterpret_cast<void *>(adj_n)
                 << " -> " << n->Name() << reinterpret_cast<void *>(n)
                 << "  via " << var->Name() << reinterpret_cast<void *>(var);
        adj_list[n].insert(adj_n);
      }
    }
  }
  return adj_list;
}

std::vector<ir::Node *> OpDFSSort(const Graph &graph) {
  auto edge_table = BuildOperationOutAdjList(graph);
  std::stack<Node *> stack;
  for (auto &ele : edge_table) {
    if (ele.first->inputs.empty()) {
      // find the input ops (those without input vars)
      stack.push(ele.first);
    } else {
      // find the ops with only persistable vars as inputs.
      bool all_persistable = true;
      for (auto *input : ele.first->inputs) {
        if (!(input->IsVar() && input->Var() && input->Var()->Persistable())) {
          all_persistable = false;
        }
      }
      if (all_persistable) {
        stack.push(ele.first);
      }
    }
  }

  std::vector<Node *> res;
  // start from the feed op and DFS
  std::unordered_set<Node *> unique_set;
  while (!stack.empty()) {
    // will start from the last feed by default.
    auto cur = stack.top();
    stack.pop();
    unique_set.insert(cur);
    res.push_back(cur);

    for (auto *op : edge_table[cur]) {
      if (!unique_set.count(op)) {
        stack.push(op);
      }
    }
  }
  return res;
}

std::vector<ir::Node *> TopologyDfsSortOperations(const Graph &graph) {
  std::vector<ir::Node *> nodes;
  std::unordered_map<Node *, int> in_degree;

  auto set_out_ops_ready = [&](Node *var) {
    for (auto *op : var->outputs) {
      --in_degree[op];
    }
  };
  // build in_degree
  for (auto *node : graph.Nodes()) {
    if (node->IsOp()) {
      in_degree[node] += node->inputs.size();
    } else if (node->IsVar() && node->inputs.empty()) {
      // put all the inputs of the whole graph ready.
      set_out_ops_ready(node);
    }
  }

  std::deque<Node *> op_queue;
  // first visit
  for (auto &node : OpDFSSort(graph)) {
    if (node->IsOp()) {
      op_queue.push_back(node);
    }
  }

  // traverse the graph
  int num_ops = op_queue.size();
  while (num_ops) {
    for (auto it = op_queue.begin(); it != op_queue.end(); it++) {
      auto *&cur_op = *it;
      if (!cur_op || in_degree[cur_op] > 0) continue;
      // visit this node
      // put all the output var of this op valid.
      for (auto *out_var : cur_op->outputs) {
        if (!out_var) continue;
        set_out_ops_ready(out_var);
      }
      VLOG(8) << "visit " << cur_op->Name();
      nodes.push_back(cur_op);

      cur_op = nullptr;
      num_ops--;
    }
  }

  return nodes;
}

C
chengduo 已提交
297
size_t GraphNum(const Graph &graph) {
D
dzhwinter 已提交
298
  std::unordered_set<ir::Node *> nodes(graph.Nodes());
C
chengduo 已提交
299 300 301 302 303
  std::unordered_set<ir::Node *> visited_nodes;
  visited_nodes.reserve(nodes.size());
  std::deque<ir::Node *> q_nodes;
  std::vector<std::unordered_set<ir::Node *>> graph_nodes;
  std::unordered_set<ir::Node *> g_nodes;
W
Wu Yi 已提交
304 305
  // q_set used to record records in the queue.
  std::unordered_set<ir::Node *> q_set;
C
chengduo 已提交
306 307
  size_t graph_count = 0;

W
Wu Yi 已提交
308 309 310 311 312 313 314 315
  auto traverse_nodes = [&visited_nodes, &q_nodes,
                         &q_set](const std::vector<ir::Node *> &nodes) {
    for (auto n : nodes) {
      if (visited_nodes.count(n) == 0 && q_set.count(n) == 0) {
        q_nodes.push_back(n);
        q_set.insert(n);
      }
    }
C
chengduo 已提交
316 317 318 319 320 321
  };

  while (visited_nodes.size() != nodes.size()) {
    if (!q_nodes.empty()) {
      auto cur_node = q_nodes.front();
      q_nodes.pop_front();
W
Wu Yi 已提交
322
      q_set.erase(cur_node);
C
chengduo 已提交
323 324 325 326 327 328 329 330 331 332 333 334 335
      visited_nodes.insert(cur_node);
      g_nodes.insert(cur_node);
      traverse_nodes(cur_node->inputs);
      traverse_nodes(cur_node->outputs);
    } else {
      ++graph_count;
      if (g_nodes.size()) {
        graph_nodes.emplace_back(g_nodes);
      }
      g_nodes.clear();
      for (auto &n : nodes) {
        if (visited_nodes.count(n) == 0) {
          q_nodes.push_back(n);
W
Wu Yi 已提交
336
          q_set.insert(n);
C
chengduo 已提交
337 338 339 340 341 342 343 344 345 346
          break;
        }
      }
    }
  }

  if (g_nodes.size()) {
    graph_nodes.emplace_back(g_nodes);
  }

C
chengduo 已提交
347 348 349 350 351 352 353 354 355
  if (FLAGS_print_sub_graph_dir.size()) {
    if (graph_nodes.size() > 1) {
      std::stringstream out;
      for (auto &g_n : graph_nodes) {
        out << "graph_nodes: " << g_n.size() << "\n";
      }
      out << "\n\n";
      for (auto &g_n : graph_nodes) {
        out << "graph_nodes: " << g_n.size();
C
chengduo 已提交
356 357 358 359 360 361 362 363 364 365 366
        for (auto &node : g_n) {
          out << "\nNode: " << node->Name() << " in [";
          for (auto &n : node->inputs) {
            out << n->Name() << ", ";
          }
          out << "], out[";
          for (auto &n : node->outputs) {
            out << n->Name() << ", ";
          }
          out << "]";
        }
C
chengduo 已提交
367
        out << "\n\n\n";
C
chengduo 已提交
368
      }
C
chengduo 已提交
369 370
      std::unique_ptr<std::ostream> fout(
          new std::ofstream(FLAGS_print_sub_graph_dir));
371 372 373 374
      PADDLE_ENFORCE_EQ(fout->good(), true,
                        platform::errors::Unavailable(
                            "Can not open file %s for printing the graph.",
                            FLAGS_print_sub_graph_dir));
C
chengduo 已提交
375
      *fout << out.str();
C
chengduo 已提交
376 377 378 379 380 381
    }
  }

  return graph_count;
}

Y
Yan Chunwei 已提交
382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404
void CleanIndividualNodes(Graph *graph) {
  std::unordered_set<Node *> nodes2rm;
  for (auto *node : graph->Nodes()) {
    if (node->inputs.empty() && node->outputs.empty()) {
      nodes2rm.insert(node);
    }
  }

  for (auto *node : nodes2rm) {
    graph->RemoveNode(node);
  }
}

std::vector<Node *> TopologyVarientSort(const Graph &graph,
                                        SortKind sort_kind) {
  switch (sort_kind) {
    case SortKind::TS:
      return framework::ir::TopologySortOperations(graph);
    default:
      return framework::ir::TopologyDfsSortOperations(graph);
  }
}

405 406
class DescOrderComparator {
 public:
407 408 409 410 411 412 413 414
  bool operator()(Node *const &n1, Node *const &n2) const {
    if (n1->DescOrder() < n2->DescOrder()) {
      return true;
    } else if (n1->DescOrder() == n2->DescOrder()) {
      return n1->id() < n2->id() ||
             (n1->id() == n2->id() && n1->ToString() < n2->ToString());
    }
    return false;
415 416 417 418
  }
};

std::vector<ir::Node *> TopologySortGraphByDescOrder(const Graph &graph) {
419 420 421 422 423 424 425 426 427 428 429
  std::map<ir::Node *, std::set<ir::Node *, DescOrderComparator>,
           DescOrderComparator>
      adj_list = BuildOperationAdjList<DescOrderComparator>(graph);
  PADDLE_ENFORCE_EQ(HasCircleInternal<DescOrderComparator>(adj_list, nullptr),
                    false, platform::errors::InvalidArgument(
                               "Generated graph shouldn't contain cycle."));
  std::unordered_set<ir::Node *> visited;
  std::vector<ir::Node *> ret;
  for (auto adj : adj_list) {
    if (visited.find(adj.first) == visited.end()) {
      SortHelper<DescOrderComparator>(adj_list, adj.first, &visited, &ret);
430 431 432
    }
  }

433
  return ret;
434 435
}

436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537
static OpDesc *ReplaceScaleLossGradOp(const Node &node, OpDesc *desc) {
  desc->SetType("fill_constant");
  desc->SetAttr(
      OpProtoAndCheckerMaker::OpRoleAttrName(),
      (static_cast<int>(OpRole::kBackward) | static_cast<int>(OpRole::kLoss)));
  desc->SetAttr("value", 1.0f);
  std::vector<std::string> output_names;
  for (auto out : node.outputs) {
    output_names.emplace_back(out->Name());
  }
  desc->SetOutput("Out", output_names);
  return desc;
}

static void GetGraphOpDesc(const std::vector<Node *> &nodes,
                           std::vector<OpDesc> *ops) {
  for (Node *n : nodes) {
    // if node is not Op, skip
    if (!n->IsOp()) continue;

    // create fill_constant op
    if (n->Name() == "scale_loss_grad") {
      ops->emplace_back();
      auto &desc = ops->back();
      ReplaceScaleLossGradOp(*n, &desc);
    } else if (n->Op()) {
      ops->emplace_back(*n->Op());
    }
    // delete no OpDesc op
  }
}

static void GraphToBlock(const Graph &graph, proto::BlockDesc *block,
                         const SortKind *sort_kind) {
  // Remove the unneeded variables after memory optimization.
  std::unordered_set<std::string> vars2remove;
  if (graph.Has(kGraphToProgramVarsToRemove)) {
    vars2remove =
        graph.Get<std::unordered_set<std::string>>(kGraphToProgramVarsToRemove);
    VLOG(2) << "graph (id: " << block->idx() << ") to program remove "
            << vars2remove.size() << " nodes";
  }

  block->clear_vars();
  std::unordered_set<std::string> visited_vars;
  for (Node *n : graph.Nodes()) {
    if (n->IsVar()) {
      if (n->Var() && visited_vars.count(n->Var()->Name()) == 0 &&
          !vars2remove.count(n->Var()->Name()) &&
          n->GetVarNodeBlockId() == graph.GetBlockId()) {
        visited_vars.insert(n->Var()->Name());
        block->add_vars()->MergeFrom(*n->Var()->Proto());
      }
    }
  }
  block->clear_ops();

  std::vector<Node *> nodes;
  if (sort_kind != nullptr) {
    // Inference Memory Optimize relays on this branch.
    nodes = TopologyVarientSort(graph, *sort_kind);
  } else {
    if (FLAGS_convert_all_blocks) {
      nodes = TopologySortGraphByDescOrder(graph);
    } else {
      nodes = TopologySortOperations(graph);
    }
  }

  std::vector<OpDesc> ops;
  GetGraphOpDesc(nodes, &ops);
  for (auto &op : ops) {
    block->add_ops()->MergeFrom(*op.Proto());
  }
}

void GraphToProgram(const Graph &graph, ProgramDesc *program,
                    const SortKind *sort_kind) {
  PADDLE_ENFORCE_EQ(graph.IsMainGraph(), true,
                    platform::errors::InvalidArgument(
                        "This graph is a sub_graph, "
                        "and can't convert to program individually"));
  PADDLE_ENFORCE_NOT_NULL(
      program,
      platform::errors::InvalidArgument(
          "program must not be nullptr when converting graph to program"));

  proto::ProgramDesc program_pb(*(program->Proto()));
  auto block = program_pb.mutable_blocks(kRootBlockIndex);
  block->set_idx(kRootBlockIndex);

  if (FLAGS_convert_all_blocks) {
    GraphToBlock(*graph.GetSubGraph(kRootBlockIndex), block, sort_kind);

    VLOG(3) << "Graph to program need convert " << graph.SubGraphsSize()
            << " sub graph";
    for (size_t idx = 0; idx < graph.SubGraphsSize(); ++idx) {
      // avoid kRootBlockIndex not 0
      if (idx == kRootBlockIndex) continue;

      block = program_pb.add_blocks();
      block->set_idx(idx);
538
      block->set_parent_idx(kRootBlockIndex);
539 540 541 542 543 544 545 546 547
      GraphToBlock(*graph.GetSubGraph(idx), block, sort_kind);
    }
  } else {
    GraphToBlock(graph, block, sort_kind);
  }

  program->CopyFrom(program_pb);
}

X
better  
Xin Pan 已提交
548 549 550
}  // namespace ir
}  // namespace framework
}  // namespace paddle