Add helper to check if node provides a graph output. (#8186)

* Add helper to check if node provides a graph output. The current approach unnecessarily creates a vector when most of the optimizers only care about a true/false response.

* Undo accidental change

* Fix a couple of issues due to copying from larger set of changes.

include/onnxruntime/core/graph/graph.h

@@ -681,6 +681,19 @@ class Graph {
     return std::find(graph_outputs_.begin(), graph_outputs_.end(), node_arg) != graph_outputs_.end();
   }
 
+  /** Returns true if one or more of the Node outputs are Graph outputs. 
+  @remarks Cheaper than calling GetNodeOutputsInGraphOutputs.
+  */
+  bool GetNodeProvidesGraphOutput(const Node& node) const {
+    auto end_outputs = graph_outputs_.cend();
+    for (auto output_def : node.OutputDefs()) {
+      if (std::find(graph_outputs_.cbegin(), end_outputs, output_def) != end_outputs) {
+        return true;
+      }
+    }
+    return false;
+  }
+
   /** Returns a vector with the indexes of the outputs of the given Node that are also Graph outputs. */
   std::vector<int> GetNodeOutputsInGraphOutputs(const Node& node) const {
     int output_idx = 0;

onnxruntime/core/graph/graph_utils.cc

@@ -323,7 +323,7 @@ bool CanRemoveNode(const Graph& graph, const Node& node, const logging::Logger&
   // This would allow removal of a node that is providing a graph output, as that output name would come from updating
   // the upstream node. This should also enable removal if CanUpdateImplicitInputNameInSubgraphs returns false.
 
-  if (!graph.GetNodeOutputsInGraphOutputs(node).empty()) {
+  if (graph.GetNodeProvidesGraphOutput(node)) {
     return false;
   }
 
@@ -762,7 +762,7 @@ bool RemoveNodesWithOneOutputBottomUp(Graph& graph, const Node& start_node) {
     // Each eligible node in the subgraph must have less than one output edge and no output should be
     // the graph output
     const Node& cur_node = *graph.GetNode(cur_node_index);
-    if (cur_node.GetOutputEdgesCount() > 1 || !graph.GetNodeOutputsInGraphOutputs(cur_node).empty()) {
+    if (cur_node.GetOutputEdgesCount() > 1 || graph.GetNodeProvidesGraphOutput(cur_node)) {
       continue;
     }
 

onnxruntime/core/optimizer/bias_dropout_fusion.cc

@@ -25,7 +25,7 @@ void FuseResidualAddIfAny(Graph& graph, const Node& dropout_node,
   // To be able to fuse the residual Add,
   // the Dropout's output must not be a graph output and
   // there must be only one consumer of the Dropout's first output.
-  if (dropout_consumers_count < 2 && graph.GetNodeOutputsInGraphOutputs(dropout_node).empty()) {
+  if (dropout_consumers_count < 2 && !graph.GetNodeProvidesGraphOutput(dropout_node)) {
     for (auto last_node_itr = dropout_node.OutputNodesBegin(); last_node_itr != dropout_node.OutputNodesEnd(); ++last_node_itr) {
       const Node& last_node = (*last_node_itr);
 
@@ -139,7 +139,7 @@ Status BiasDropoutFusion::ApplyImpl(Graph& graph, bool& modified, int graph_leve
       continue;
     }
 
-    if (!graph.GetNodeOutputsInGraphOutputs(node).empty()) {
+    if (graph.GetNodeProvidesGraphOutput(node)) {
       continue;
     }
 

onnxruntime/core/optimizer/bias_gelu_fusion.cc

@@ -76,7 +76,7 @@ Status BiasGeluFusion::ApplyImpl(Graph& graph, bool& modified, int graph_level,
       continue;
     }
 
-    if (!graph.GetNodeOutputsInGraphOutputs(node).empty()) {
+    if (graph.GetNodeProvidesGraphOutput(node)) {
       continue;
     }
 

onnxruntime/core/optimizer/conv_activation_fusion.cc

@@ -96,12 +96,12 @@ Status ConvActivationFusion::ApplyImpl(Graph& graph, bool& modified, int graph_l
       continue;
     }
 
-    if (!graph.GetNodeOutputsInGraphOutputs(*node).empty()) {
+    if (graph.GetNodeProvidesGraphOutput(*node)) {
       continue;
     }
 
     if (node->GetExecutionProviderType() == onnxruntime::kCudaExecutionProvider) {
-      if (node->InputDefs()[0]->TypeAsProto()->tensor_type().elem_type() != 
+      if (node->InputDefs()[0]->TypeAsProto()->tensor_type().elem_type() !=
           ONNX_NAMESPACE::TensorProto_DataType_FLOAT) {
         continue;
       }
@@ -125,7 +125,7 @@ Status ConvActivationFusion::ApplyImpl(Graph& graph, bool& modified, int graph_l
         if (last_node.GetExecutionProviderType() != node->GetExecutionProviderType()) {
           continue;
         }
-        if (graph_utils::IsSupportedOptypeVersionAndDomain(last_node, "Relu", {6, 13, 14}) && 
+        if (graph_utils::IsSupportedOptypeVersionAndDomain(last_node, "Relu", {6, 13, 14}) &&
             next_node.GetOutputEdgesCount() == 1) {
           Node& conv_node = *node;
           Node& add_node = *graph.GetNode(next_node.Index());

onnxruntime/core/optimizer/conv_add_fusion.cc

@@ -125,7 +125,7 @@ bool ConvAddFusion::SatisfyCondition(const Graph& graph, const Node& node, const
     return false;
   }
 
-  if (!graph.GetNodeOutputsInGraphOutputs(node).empty()) {
+  if (graph.GetNodeProvidesGraphOutput(node)) {
     return false;
   }
 

onnxruntime/core/optimizer/conv_bn_fusion.cc

@@ -177,7 +177,7 @@ bool ConvBNFusion::SatisfyCondition(const Graph& graph, const Node& node, const
     }
   }
 
-  if (!graph.GetNodeOutputsInGraphOutputs(node).empty()) {
+  if (graph.GetNodeProvidesGraphOutput(node)) {
     return false;
   }
 

onnxruntime/core/optimizer/conv_mul_fusion.cc

@@ -133,7 +133,7 @@ bool ConvMulFusion::SatisfyCondition(const Graph& graph, const Node& node, const
     return false;
   }
 
-  if (!graph.GetNodeOutputsInGraphOutputs(node).empty()) {
+  if (graph.GetNodeProvidesGraphOutput(node)) {
     return false;
   }
 

onnxruntime/core/optimizer/div_mul_fusion.cc

@@ -74,7 +74,7 @@ bool DivMulFusion::SatisfyCondition(const Graph& graph, const Node& node, const
       return false;
   }
 
-  if (!graph.GetNodeOutputsInGraphOutputs(node).empty()) {
+  if (graph.GetNodeProvidesGraphOutput(node)) {
     return false;
   }
 

onnxruntime/core/optimizer/fast_gelu_fusion.cc

@@ -31,7 +31,7 @@ static bool CheckNode(Graph& graph, const Node& node, const std::string& op_name
          node.GetExecutionProviderType() == provider &&
          IsSupportedDataType(node) &&
          (!require_single_output || node.GetOutputEdgesCount() == 1) &&
-         graph.GetNodeOutputsInGraphOutputs(node).empty();
+         !graph.GetNodeProvidesGraphOutput(node);
 }
 
 MatchResult FastGeluFusion::CheckFirstFormula(Graph& graph, Node& mul1_node,
@@ -146,8 +146,8 @@ MatchResult FastGeluFusion::CheckSecondFormula(Graph& graph, Node& pow1_node,
   if (p_cast1_node != nullptr) {
     Node& cast1_node = *graph.GetNode(p_cast1_node->Index());
     // this is fused Cast node, so expect 2 output edges
-    if (!CheckNode(graph, cast1_node, "Cast", {9, 13}, pow1_node.GetExecutionProviderType(), false)  ||
-      cast1_node.GetOutputEdgesCount() != 2){
+    if (!CheckNode(graph, cast1_node, "Cast", {9, 13}, pow1_node.GetExecutionProviderType(), false) ||
+        cast1_node.GetOutputEdgesCount() != 2) {
       return matchResult;
     }
     const Node* p_pow_node = graph_utils::FirstChildByType(cast1_node, "Pow");
@@ -242,7 +242,7 @@ Status FastGeluFusion::ApplyImpl(Graph& graph, bool& modified, int graph_level,
     // if this is second formula and if pow node has Cast parent, expect mul5_node has Cast parent as well
     NodeArg* cast_input_arg = nullptr;
     if (second_formula) {
-      const Node* p_cast1_node = graph_utils::FirstParentByType(node, "Cast");   
+      const Node* p_cast1_node = graph_utils::FirstParentByType(node, "Cast");
       if (p_cast1_node != nullptr) {
         // we've done the node check in second formula for pow node
         Node& cast1_node = *graph.GetNode(p_cast1_node->Index());
@@ -254,11 +254,11 @@ Status FastGeluFusion::ApplyImpl(Graph& graph, bool& modified, int graph_level,
         Node& cast3_node = *graph.GetNode(p_cast3_node->Index());
         if (!CheckNode(graph, cast3_node, "Cast", {9, 13}, node.GetExecutionProviderType(), true)) {
           continue;
-        } 
+        }
         // overwrite and continue as usual
-        p_mul5_input_node = graph_utils::FirstParentByType(cast3_node, "Mul");        
-        nodes_to_fuse.push_back(cast3_node);   
-        // keep cast1_node for reuse, its output edges will be adjusted in FinalizeNodeFusion()                         
+        p_mul5_input_node = graph_utils::FirstParentByType(cast3_node, "Mul");
+        nodes_to_fuse.push_back(cast3_node);
+        // keep cast1_node for reuse, its output edges will be adjusted in FinalizeNodeFusion()
       }
     }
 
@@ -275,8 +275,8 @@ Status FastGeluFusion::ApplyImpl(Graph& graph, bool& modified, int graph_level,
       }
     }
 
-    if (input_index == -1)  continue;
-    // check same parent for both mul6 and pow, with or without cast 
+    if (input_index == -1) continue;
+    // check same parent for both mul6 and pow, with or without cast
     if (cast_input_arg != nullptr) {
       if (mul6_node.InputDefs()[(input_index + 1) % 2]->Name() != cast_input_arg->Name())
         continue;

onnxruntime/core/optimizer/gemm_activation_fusion.cc

@@ -61,7 +61,7 @@ Status GemmActivationFusion::ApplyImpl(Graph& graph, bool& modified, int graph_l
       continue;
     }
 
-    if (!graph.GetNodeOutputsInGraphOutputs(node).empty()) {
+    if (graph.GetNodeProvidesGraphOutput(node)) {
       continue;
     }
 

onnxruntime/core/optimizer/gemm_transpose_fusion.cc

@@ -42,8 +42,8 @@ Status GemmTransposeFusion::Apply(Graph& graph, Node& node, RewriteRuleEffect& m
   nodes_to_remove.push_back(gemm_node);
 
   // check if output node is Transpose
-  if (output_node_ptr != gemm_node.OutputNodesEnd() && 
-      gemm_node.InputDefs().size() <= 2 && // C is missing
+  if (output_node_ptr != gemm_node.OutputNodesEnd() &&
+      gemm_node.InputDefs().size() <= 2 &&  // C is missing
       output_node_ptr->OpType() == "Transpose") {
     Node& output_node = *graph.GetNode(output_node_ptr->Index());
     // (AB)' = B'A' : reverse the inputs
@@ -83,7 +83,7 @@ bool GemmTransposeFusion::SatisfyCondition(const Graph& graph, const Node& node,
   for (auto node_it = node.InputNodesBegin(); node_it != node.InputNodesEnd(); ++node_it) {
     if (graph_utils::IsSupportedOptypeVersionAndDomain(*node_it, "Transpose", {1, 13}) &&
         node_it->GetOutputEdgesCount() == 1 &&
-        graph.GetNodeOutputsInGraphOutputs(*node_it).empty() &&
+        !graph.GetNodeProvidesGraphOutput(*node_it) &&
         // Make sure the two nodes do not span execution providers.
         node_it->GetExecutionProviderType() == node.GetExecutionProviderType()) {
       return true;
@@ -94,7 +94,7 @@ bool GemmTransposeFusion::SatisfyCondition(const Graph& graph, const Node& node,
   // by the rule (AB)' = B'A' provided that C is missing
   // Supported for Opset >=11 as earlier opsets have C as a required input
   if (!graph_utils::IsSupportedOptypeVersionAndDomain(node, "Gemm", {11, 13}) ||
-      !graph.GetNodeOutputsInGraphOutputs(node).empty() ||
+      graph.GetNodeProvidesGraphOutput(node) ||
       // verify that C is missing
       node.InputDefs().size() > 2) {
     return false;

onnxruntime/core/optimizer/identity_elimination.cc

@@ -40,7 +40,7 @@ namespace onnxruntime {
       X (def0/arg0) ---> Identity ---> Y
  */
 Status EliminateIdentity::Apply(Graph& graph, Node& node, RewriteRuleEffect& rule_effect, const logging::Logger&) const {
-  if (graph.GetNodeOutputsInGraphOutputs(node).empty()) {
+  if (!graph.GetNodeProvidesGraphOutput(node)) {
     if (graph_utils::RemoveNode(graph, node)) {
       rule_effect = RewriteRuleEffect::kRemovedCurrentNode;
     }
@@ -65,7 +65,7 @@ bool EliminateIdentity::SatisfyCondition(const Graph& graph, const Node& node, c
     return true;
   }
 
-  bool node_output_is_graph_output = !graph.GetNodeOutputsInGraphOutputs(node).empty();
+  bool node_output_is_graph_output = graph.GetNodeProvidesGraphOutput(node);
 
   // relax the condition if Identity is connecting to graph output
   if (node.GetOutputEdgesCount() != 0 ||

onnxruntime/core/optimizer/insert_cast_transformer.cc

@@ -94,7 +94,7 @@ static bool IsIsolatedFp16NodeOnCpu(const onnxruntime::Node& node, onnxruntime::
   //    and is assigned to the CPU EP (we have fp32 implementations of all kernels so forcing to fp32 is safe)
   if (node.GetInputEdgesCount() > 0 &&
       !node.ContainsSubgraph() &&
-      graph.GetNodeOutputsInGraphOutputs(node).empty() &&
+      !graph.GetNodeProvidesGraphOutput(node) &&
       node.GetExecutionProviderType() == kCpuExecutionProvider) {
     do {
       // find the number of fp16 inputs as we need to make sure they're all coming from nodes that will be cast

onnxruntime/core/optimizer/isinf_reducesum_fusion.cc

@@ -35,7 +35,7 @@ Status IsInfReduceSumFusion::ApplyImpl(Graph& graph, bool& modified, int graph_l
 
     if (!graph_utils::IsSupportedOptypeVersionAndDomain(isinf_node, "IsInf", {10}) ||
         isinf_node.GetOutputEdgesCount() != 1 ||
-        !graph.GetNodeOutputsInGraphOutputs(isinf_node).empty()) {
+        graph.GetNodeProvidesGraphOutput(isinf_node)) {
       continue;
     }
 
@@ -67,7 +67,7 @@ Status IsInfReduceSumFusion::ApplyImpl(Graph& graph, bool& modified, int graph_l
     Node& cast2_node = *graph.GetNode(cast2_node_itr->Index());
     if (!graph_utils::IsSupportedOptypeVersionAndDomain(cast2_node, "Cast", {9, 13}) ||
         cast2_node.GetOutputEdgesCount() != 1 ||
-        !graph.GetNodeOutputsInGraphOutputs(cast2_node).empty()) {
+        graph.GetNodeProvidesGraphOutput(cast2_node)) {
       continue;
     }
     nodes_to_remove.push_back(cast2_node);
@@ -80,7 +80,7 @@ Status IsInfReduceSumFusion::ApplyImpl(Graph& graph, bool& modified, int graph_l
     Node& reduce_sum_node = *graph.GetNode(reduce_sum_node_itr->Index());
     if (!graph_utils::IsSupportedOptypeVersionAndDomain(reduce_sum_node, "ReduceSum", {1, 11, 13}) ||
         reduce_sum_node.GetOutputEdgesCount() != 1 ||
-        !graph.GetNodeOutputsInGraphOutputs(reduce_sum_node).empty()) {
+        graph.GetNodeProvidesGraphOutput(reduce_sum_node)) {
       continue;
     }
     nodes_to_remove.push_back(reduce_sum_node);

onnxruntime/core/optimizer/layer_norm_fusion.cc

@@ -79,7 +79,7 @@ Status LayerNormFusion::ApplyImpl(Graph& graph, bool& modified, int graph_level,
     if (!graph_utils::IsSupportedOptypeVersionAndDomain(reduce_mean_node, "ReduceMean", {1, 11, 13}) ||
         !graph_utils::IsSupportedProvider(reduce_mean_node, GetCompatibleExecutionProviders()) ||
         (reduce_mean_node.GetOutputEdgesCount() != 1 && reduce_mean_node.GetOutputEdgesCount() != 2) ||
-        !graph.GetNodeOutputsInGraphOutputs(reduce_mean_node).empty() ||
+        graph.GetNodeProvidesGraphOutput(reduce_mean_node) ||
         !IsSupportedDataType(reduce_mean_node)) {
       continue;
     }
@@ -377,7 +377,7 @@ Status SimplifiedLayerNormFusion::ApplyImpl(Graph& graph, bool& modified, int gr
     if (!graph_utils::IsSupportedOptypeVersionAndDomain(pow_node, "Pow", {7, 12, 13}) ||
         !graph_utils::IsSupportedProvider(pow_node, GetCompatibleExecutionProviders()) ||
         !optimizer_utils::CheckOutputEdges(graph, pow_node, 1) ||
-        !graph.GetNodeOutputsInGraphOutputs(pow_node).empty() ||
+        graph.GetNodeProvidesGraphOutput(pow_node) ||
         !IsSupportedDataType(pow_node)) {
       continue;
     }

onnxruntime/core/optimizer/matmul_add_fusion.cc

@@ -30,7 +30,7 @@ Status MatMulAddFusion::ApplyImpl(Graph& graph, bool& modified, int graph_level,
       continue;
     }
 
-    if (!graph.GetNodeOutputsInGraphOutputs(node).empty()) {
+    if (graph.GetNodeProvidesGraphOutput(node)) {
       continue;
     }
 
@@ -89,7 +89,7 @@ Status MatMulAddFusion::ApplyImpl(Graph& graph, bool& modified, int graph_level,
     // valid bias_shapes are (N) or (1, N) or (M, 1) or (M, N) as
     // GEMM only supports unidirectional broadcast on the bias input C
     if (!gemm_input_defs.back()->Shape()) {
-        continue;
+      continue;
     }
     const auto& bias_shape = *gemm_input_defs.back()->Shape();
     const auto& M = matmul_output.Shape()->dim()[0];

onnxruntime/core/optimizer/matmul_transpose_fusion.cc

@@ -39,7 +39,7 @@ static Node* GetTransposeNodeFromOutput(Graph& graph, NodeArg& node_arg) {
   }
 
   // if the node has Graph output, skip it too
-  if (!graph.GetNodeOutputsInGraphOutputs(*trans_node).empty()) {
+  if (graph.GetNodeProvidesGraphOutput(*trans_node)) {
     return nullptr;
   }
 
@@ -117,9 +117,8 @@ static size_t UpdateConsumerCount(Graph& graph, NodeArg* target, std::unordered_
 *                              V
 */
 static Node* ReorderCastAndTranspose(Graph& graph, Node* cast,
-                                    std::unordered_map<NodeArg*, size_t>& consumer_count,
-                                    std::deque<onnxruntime::NodeIndex>& removed_nodes) {
-
+                                     std::unordered_map<NodeArg*, size_t>& consumer_count,
+                                     std::deque<onnxruntime::NodeIndex>& removed_nodes) {
   ORT_ENFORCE(cast != nullptr);
   auto transpose = GetTransposeNodeFromOutput(graph, *cast->MutableInputDefs()[0]);
   if (transpose == nullptr) {
@@ -138,18 +137,18 @@ static Node* ReorderCastAndTranspose(Graph& graph, Node* cast,
   new_cast_output_type_proto.mutable_tensor_type()->set_elem_type(element_type);
   auto& new_cast_output = graph.GetOrCreateNodeArg(cast_output->Name() + "_transformed", &new_cast_output_type_proto);
 
-  const std::vector<NodeArg*> new_cast_input_defs {transpose_input};
-  const std::vector<NodeArg*> new_cast_output_defs {&new_cast_output};
+  const std::vector<NodeArg*> new_cast_input_defs{transpose_input};
+  const std::vector<NodeArg*> new_cast_output_defs{&new_cast_output};
   const std::vector<NodeArg*> new_transpose_input_defs = {&new_cast_output};
   const std::vector<NodeArg*> new_transpose_output_defs = {cast_output};
 
-  (void) graph.AddNode(graph.GenerateNodeName(cast->Name() + "_transformed"),
-                                 cast->OpType(),
-                                 "Created a new Cast node to interchange Cast and Transpose nodes",
-                                 new_cast_input_defs,
-                                 new_cast_output_defs,
-                                 &cast->GetAttributes(),
-                                 cast->Domain());
+  (void)graph.AddNode(graph.GenerateNodeName(cast->Name() + "_transformed"),
+                      cast->OpType(),
+                      "Created a new Cast node to interchange Cast and Transpose nodes",
+                      new_cast_input_defs,
+                      new_cast_output_defs,
+                      &cast->GetAttributes(),
+                      cast->Domain());
 
   Node& new_transpose = graph.AddNode(graph.GenerateNodeName(transpose->Name() + "_transformed"),
                                       transpose->OpType(),
@@ -169,8 +168,7 @@ static Node* ReorderCastAndTranspose(Graph& graph, Node* cast,
 }
 
 // Check whether the element_type is an allowed FusedMatMul data type or not.
-static bool IsAllowedFusedMatMulDataType(ONNX_NAMESPACE::TensorProto_DataType element_type)
-{
+static bool IsAllowedFusedMatMulDataType(ONNX_NAMESPACE::TensorProto_DataType element_type) {
   return element_type == ONNX_NAMESPACE::TensorProto_DataType_FLOAT ||
          element_type == ONNX_NAMESPACE::TensorProto_DataType_FLOAT16 ||
          element_type == ONNX_NAMESPACE::TensorProto_DataType_DOUBLE ||

onnxruntime/core/optimizer/nchwc_transformer.cc

@@ -164,7 +164,7 @@ size_t NchwcTransformerImpl::RemoveOutputEdges(Node& node) {
   }
   // Bias the edge count to handle the case of a node that produces a graph
   // output.
-  if (!graph_.GetNodeOutputsInGraphOutputs(node).empty()) {
+  if (graph_.GetNodeProvidesGraphOutput(node)) {
     output_edges_count++;
   }
   return output_edges_count;
@@ -1145,7 +1145,7 @@ void NchwcTransformerImpl::TrackTransposeFromNhwc(Node& node) {
 
   // Verify that the node does not produce a graph output and produces output
   // for a single node.
-  if (!graph_.GetNodeOutputsInGraphOutputs(node).empty() || node.GetOutputEdgesCount() != 1) {
+  if (graph_.GetNodeProvidesGraphOutput(node) || node.GetOutputEdgesCount() != 1) {
     return;
   }
 

onnxruntime/core/optimizer/qdq_transformer/qdq_propagation.cc

@@ -26,7 +26,7 @@ static bool TryCancelOutDQQPair(Graph& graph, Node& dq_node, Node& q_node) {
   // check if dq_node has only one output edge and,
   // dq_node and q_node output are not graph outputs
   if (!optimizer_utils::CheckOutputEdges(graph, dq_node, 1) ||
-      !graph.GetNodeOutputsInGraphOutputs(q_node).empty()) {
+      graph.GetNodeProvidesGraphOutput(q_node)) {
     return false;
   }
 

onnxruntime/core/optimizer/skip_layer_norm_fusion.cc

@@ -173,8 +173,8 @@ Status SkipLayerNormFusion::ApplyImpl(Graph& graph, bool& modified, int graph_le
 
       if (CheckFirstAdd(*p_add1, ln_node.GetExecutionProviderType()) &&
           CheckSecondAdd(graph, *p_add2, ln_node.GetExecutionProviderType()) &&
-          graph.GetNodeOutputsInGraphOutputs(*p_add1).empty() &&
-          graph.GetNodeOutputsInGraphOutputs(*p_add2).empty()) {
+          !graph.GetNodeProvidesGraphOutput(*p_add1) &&
+          !graph.GetNodeProvidesGraphOutput(*p_add2)) {
         matched_format = Format::Format1;
       }
     }
@@ -191,8 +191,8 @@ Status SkipLayerNormFusion::ApplyImpl(Graph& graph, bool& modified, int graph_le
 
         if (CheckFirstAdd(*p_add1, ln_node.GetExecutionProviderType()) &&
             CheckSecondAdd(graph, *p_add2, ln_node.GetExecutionProviderType()) &&
-            graph.GetNodeOutputsInGraphOutputs(*p_add1).empty() &&
-            graph.GetNodeOutputsInGraphOutputs(*p_add2).empty()) {
+            !graph.GetNodeProvidesGraphOutput(*p_add1) &&
+            !graph.GetNodeProvidesGraphOutput(*p_add2)) {
           matched_format = Format::Format2;
         }
       }

onnxruntime/core/optimizer/utils.cc

@@ -267,7 +267,7 @@ int32_t IndexOfNodeOutput(const Node& node, const NodeArg& node_arg) {
 }
 
 bool CheckOutputEdges(const Graph& graph, const Node& node, size_t expected_output_edges) {
-  if (!graph.GetNodeOutputsInGraphOutputs(node).empty()) {
+  if (graph.GetNodeProvidesGraphOutput(node)) {
     return false;
   }