Bugfix reduce actor act cnt (#1013)

* bugfix: fix node2component_id bug according to new depth definition

* only take account of ActNode with output


Former-commit-id: 4a0f6219

oneflow/core/graph/act_graph.cpp

@@ -445,15 +445,6 @@ ActGraph::ActGraph(const Plan& plan, std::unique_ptr<std::list<ActEvent>>&& act_
   InitEdges();
   InitDepth();
   InitTaskId2TaskProto();
-  InitActorStatistics();
-}
-
-void ActGraph::InitActorStatistics() {
-  for (const ActEvent& act_event : *act_events_) {
-    int64_t actor_id = act_event.actor_id();
-    ++actor_id2act_cnt_[actor_id];
-    actor_id2total_act_time_[actor_id] += Duration4ActEvent(act_event);
-  }
 }
 
 void ActGraph::ForEachDepthRangeRegstUids(

oneflow/core/graph/act_graph.h

@@ -73,10 +73,6 @@ class ActGraph final : public Graph<ActNode, ActEdge> {
   const TaskProto& GetTaskProto(int64_t actor_id) const {
     return *task_id2task_proto_.at(actor_id);
   }
-  const HashMap<int64_t, int64_t>& actor_id2act_cnt() const { return actor_id2act_cnt_; }
-  const HashMap<int64_t, double>& actor_id2total_act_time() const {
-    return actor_id2total_act_time_;
-  }
   const std::list<const ActNode*>& Nodes4Depth(int64_t depth) const {
     return depth2nodes_.at(depth);
   }
@@ -94,7 +90,6 @@ class ActGraph final : public Graph<ActNode, ActEdge> {
   void InitEdges();
   void InitDepth();
   void InitTaskId2TaskProto();
-  void InitActorStatistics();
   void ForEachDepthRangeRegstUids(
       const std::function<void(const Range& range, const std::list<std::string>& regst_uids)>&
           Handler) const;
@@ -110,8 +105,6 @@ class ActGraph final : public Graph<ActNode, ActEdge> {
   HashMap<std::string, ActNode*> regst_uid2producer_node_;
   HashMap<std::string, std::list<const ActNode*>> regst_uid2consumer_nodes_;
   HashMap<int64_t, std::list<const ActNode*>> depth2nodes_;
-  HashMap<int64_t, int64_t> actor_id2act_cnt_;
-  HashMap<int64_t, double> actor_id2total_act_time_;
 };
 
 }  // namespace oneflow

oneflow/core/job/improver.cpp

@@ -230,20 +230,27 @@ double FormalDuration4ExperimentalDuration(TaskType task_type, double duration,
 
 double CalcBaseII(const ActGraph& act_graph) {
   int64_t max_act_cnt = 0;
-  for (const auto& pair : act_graph.actor_id2act_cnt()) {
-    if (max_act_cnt < pair.second) { max_act_cnt = pair.second; }
-  }
+  HashMap<int64_t, int64_t> actor_id2outputed_act_cnt;
+  act_graph.ForEachNode([&](const ActNode* act_node) {
+    int64_t actor_id = act_node->actor_id();
+    if (!act_node->out_edges().empty()) {
+      ++actor_id2outputed_act_cnt[actor_id];
+      max_act_cnt = std::max(max_act_cnt, actor_id2outputed_act_cnt[actor_id]);
+    }
+  });
   HashMap<int64_t, double> actor_id2act_frequency;
-  for (const auto& pair : act_graph.actor_id2act_cnt()) {
+  for (const auto& pair : actor_id2outputed_act_cnt) {
     actor_id2act_frequency[pair.first] = 1.0 * pair.second / max_act_cnt;
   }
   HashMap<int64_t, double> stream_id2total_calc_time;
   act_graph.ForEachNode([&](const ActNode* act_node) {
-    int64_t stream_id = act_node->act_event().work_stream_id();
     int64_t actor_id = act_node->actor_id();
+    auto frequence_it = actor_id2act_frequency.find(actor_id);
+    if (frequence_it == actor_id2act_frequency.end()) { return; }
+    int64_t stream_id = act_node->act_event().work_stream_id();
     TaskType task_type = act_graph.GetTaskProto(actor_id).task_type();
-    stream_id2total_calc_time[stream_id] += FormalDuration4ExperimentalDuration(
-        task_type, act_node->Duration(), actor_id2act_frequency.at(actor_id));
+    stream_id2total_calc_time[stream_id] +=
+        FormalDuration4ExperimentalDuration(task_type, act_node->Duration(), frequence_it->second);
   });
   double base_ii = 0;
   for (const auto& pair : stream_id2total_calc_time) {