introduce a master switch to turn on and off profiler

mindspore/ccsrc/minddata/dataset/engine/perf/connector_throughput.h

@@ -68,7 +68,7 @@ class ConnectorThroughput : public Sampling {
   // @return Status The status code returned
   Status SaveToFile() override;
 
-  Status Init(const std::string &dir_path, const std::string &device_id);
+  Status Init(const std::string &dir_path, const std::string &device_id) override;
 
   json ParseOpInfo(const DatasetOp &node, const std::vector<double> &thr);
 

mindspore/ccsrc/minddata/dataset/engine/perf/profiling.cc

@@ -31,17 +31,11 @@ namespace mindspore {
 namespace dataset {
 
 // Constructor
-ProfilingManager::ProfilingManager(ExecutionTree *tree) : tree_(tree) {
+ProfilingManager::ProfilingManager(ExecutionTree *tree) : tree_(tree), enabled_(true) {
   perf_monitor_ = std::make_unique<Monitor>(tree_);
 }
 
-bool ProfilingManager::IsProfilingEnable() const {
-  auto profiling = common::GetEnv("PROFILING_MODE");
-  if (profiling.empty() || profiling != "true") {
-    return false;
-  }
-  return true;
-}
+bool ProfilingManager::IsProfilingEnable() const { return common::GetEnv("PROFILING_MODE") == "true" && enabled_; }
 
 Status ProfilingManager::Initialize() {
   // Register nodes based on config
@@ -64,7 +58,7 @@ Status ProfilingManager::Initialize() {
 #endif
   dir_path_ = real_path;
 
-  // If DEVICE_ID is not set,defult value is 0
+  // If DEVICE_ID is not set, default value is 0
   device_id_ = common::GetEnv("DEVICE_ID");
   if (device_id_.empty()) {
     device_id_ = "0";

mindspore/ccsrc/minddata/dataset/engine/perf/profiling.h

@@ -102,9 +102,14 @@ class ProfilingManager {
   // @return Status The status code returned
   Status GetTracingNode(const std::string &name, std::shared_ptr<Tracing> *node);
 
-  // If profiling is enabled.
+  // return true if env variable has profiling enabled and enabled_ is set to true.
   bool IsProfilingEnable() const;
 
+  // Calling this would disable Profiling functionality for the entire duration of ExecutionTree. It cannot be
+  // re-enabled. Each execution_tree is associated with a unique profiling_manager which will start when tree is
+  // launched. This is the master off switch, once called, it won't start profiler even if env variable says so.
+  void DisableProfiling() { enabled_ = false; }
+
   const std::unordered_map<std::string, std::shared_ptr<Sampling>> &GetSamplingNodes() { return sampling_nodes_; }
 
   // Launch monitoring thread.
@@ -114,6 +119,7 @@ class ProfilingManager {
 
  private:
   std::unique_ptr<Monitor> perf_monitor_;
+  bool enabled_;
   std::unordered_map<std::string, std::shared_ptr<Tracing>> tracing_nodes_;
 
   std::unordered_map<std::string, std::shared_ptr<Sampling>> sampling_nodes_;
@@ -128,9 +134,9 @@ class ProfilingManager {
   // @return Status The status code returned
   Status RegisterSamplingNode(std::shared_ptr<Sampling> node);
 
-  ExecutionTree *tree_ = nullptr;  // ExecutionTree pointer
-  std::string dir_path_;           // where to create profiling file
-  std::string device_id_;          // used when create profiling file,filename_deviceid.suffix
+  ExecutionTree *tree_;    // ExecutionTree pointer
+  std::string dir_path_;   // where to create profiling file
+  std::string device_id_;  // used when create profiling file,filename_device_id.suffix
 };
 
 enum ProfilingType { TIME, CONNECTOR_DEPTH };

mindspore/ccsrc/minddata/dataset/engine/tree_adapter.cc

@@ -126,7 +126,8 @@ Status TreeAdapter::BuildExecutionTreeRecur(std::shared_ptr<DatasetNode> ir, std
 Status TreeAdapter::Build(std::shared_ptr<DatasetNode> root_ir, int32_t num_epochs) {
   // This will evolve in the long run
   tree_ = std::make_unique<ExecutionTree>();
-
+  // disable profiling if this is only a getter pass
+  if (usage_ == kDeGetter) tree_->GetProfilingManager()->DisableProfiling();
   // Build the Execution tree from the child of the IR root node, which represent the root of the input IR tree
   std::shared_ptr<DatasetOp> root_op;
   RETURN_IF_NOT_OK(BuildExecutionTreeRecur(root_ir->Children()[0], &root_op));

tests/ut/python/dataset/test_profiling.py

@@ -40,6 +40,12 @@ def test_profiling_simple_pipeline():
     data1 = ds.GeneratorDataset(source, ["data"])
     data1 = data1.shuffle(64)
     data1 = data1.batch(32)
+    # try output shape type and dataset size and make sure no profiling file is generated
+    assert data1.output_shapes() == [[32, 1]]
+    assert [str(tp) for tp in data1.output_types()] == ["int64"]
+    assert data1.get_dataset_size() == 32
+    assert os.path.exists(PIPELINE_FILE) is False
+    assert os.path.exists(DATASET_ITERATOR_FILE) is False
 
     for _ in data1:
         pass
@@ -92,7 +98,7 @@ def test_profiling_complex_pipeline():
     del os.environ['MINDDATA_PROFILING_DIR']
 
 
-def test_profiling_sampling_iterval():
+def test_profiling_sampling_interval():
     """
     Test non-default monitor sampling interval
     """
@@ -126,4 +132,4 @@ def test_profiling_sampling_iterval():
 if __name__ == "__main__":
     test_profiling_simple_pipeline()
     test_profiling_complex_pipeline()
-    test_profiling_sampling_iterval()
+    test_profiling_sampling_interval()