aboutsummaryrefslogtreecommitdiffhomepage
path: root/tensorflow/core/common_runtime/step_stats_collector.cc
diff options
context:
space:
mode:
authorGravatar Yuan Yu <yuanbyu@google.com>2016-05-18 19:25:50 -0800
committerGravatar TensorFlower Gardener <gardener@tensorflow.org>2016-05-18 20:33:03 -0700
commitbe8a3e2cf7738442f2b03c35b2ef9113b178855c (patch)
tree40e3825bd309df0971e76bbe6d6dee7fda4dfba8 /tensorflow/core/common_runtime/step_stats_collector.cc
parentc4119befd020d41be1d753267ed238fd309931df (diff)
Add the ability to return the cost model to the client as part of run metadata.
Change: 122696039
Diffstat (limited to 'tensorflow/core/common_runtime/step_stats_collector.cc')
-rw-r--r--tensorflow/core/common_runtime/step_stats_collector.cc32
1 files changed, 16 insertions, 16 deletions
diff --git a/tensorflow/core/common_runtime/step_stats_collector.cc b/tensorflow/core/common_runtime/step_stats_collector.cc
index 41ffcd1e1f..846814a1b3 100644
--- a/tensorflow/core/common_runtime/step_stats_collector.cc
+++ b/tensorflow/core/common_runtime/step_stats_collector.cc
@@ -24,24 +24,24 @@ StepStatsCollector::StepStatsCollector(StepStats* ss,
CostModelManager* cost_model_manager)
: step_stats_(ss), cost_model_manager_(cost_model_manager) {}
-void StepStatsCollector::UpdateCostModel(const NodeExecStats* nt,
- const Graph* graph, const Node* node) {
+void StepStatsCollector::UpdateCostModelNode(const NodeExecStats* nt,
+ const Graph* graph,
+ const Node* node) {
mutex_lock l(mu_);
- if (cost_model_manager_ == nullptr) {
- return;
- }
- CostModel* cm = cost_model_manager_->FindOrCreateCostModel(graph);
- cm->RecordMaxExecutionTime(node, Microseconds(nt->op_end_rel_micros()));
+ if (cost_model_manager_ != nullptr) {
+ CostModel* cm = cost_model_manager_->FindOrCreateCostModel(graph);
+ cm->RecordMaxExecutionTime(node, Microseconds(nt->op_end_rel_micros()));
- for (int i = 0; i < nt->output_size(); ++i) {
- cm->RecordMaxMemSize(node, i, Bytes(nt->output(i)
- .tensor_description()
- .allocation_description()
- .allocated_bytes()));
- cm->RecordAliases(node, i, nt->output(i)
- .tensor_description()
- .allocation_description()
- .allocation_id());
+ for (int i = 0; i < nt->output_size(); ++i) {
+ cm->RecordMaxMemorySize(node, i, Bytes(nt->output(i)
+ .tensor_description()
+ .allocation_description()
+ .allocated_bytes()));
+ cm->RecordAllocationId(node, i, nt->output(i)
+ .tensor_description()
+ .allocation_description()
+ .allocation_id());
+ }
}
}