Let execution fall back to CPU EP if Compile of a partition on current EP fails (#6580)

* Let exccution fall back to CPU EP if compile of a partition fails

* Removed debugging logs

* Addressed CR comments

include/onnxruntime/core/graph/graph.h

@@ -921,11 +921,18 @@ class Graph {
   @returns Node with fused subgraph.
   @remarks As a new Graph instance for the fused nodes is not created, a GraphViewer can be constructed with the
            IndexedSubGraph information to provide a view of the subgraph. The original nodes are left in place
-           while this is in use. 
+           while this is in use.
 		   Call FinalizeFuseSubGraph to remove them once the fused replacement node is fully created.
   */
   Node& BeginFuseSubGraph(const IndexedSubGraph& sub_graph, const std::string& fused_node_name);
 
+  /**
+  If we have BeginFuseSubGraph, but somehow hit errors, such as Compile of an EP failed on thesub_graph.
+  We can call CancelFuseSubGraph to undo the changes of BeginFuseSubGraph
+  @param fused_node The fused node and it's function body to be removed from the graph
+  */
+  void CancelFuseSubGraph(const Node& fused_node);
+
   void FinalizeFuseSubGraph(const IndexedSubGraph& sub_graph, Node& fused_node);
 #endif
 

onnxruntime/core/framework/graph_partitioner.cc

@@ -448,42 +448,43 @@ static Status PartitionOrtFormatModelImpl(Graph& graph, FuncManager& func_mgr,
     nodes_and_viewers.push_back(IExecutionProvider::FusedNodeAndGraph{fused_node, *viewers.back()});
   }
 
-  std::vector<NodeComputeInfo> node_compute_funcs;
-  node_compute_funcs.reserve(nodes_and_viewers.size());
-
-  ORT_RETURN_IF_ERROR(current_ep.Compile(nodes_and_viewers, node_compute_funcs));
-
-  if (node_compute_funcs.size() != nodes_and_viewers.size()) {
-    return ORT_MAKE_STATUS(ONNXRUNTIME, FAIL, type, " did not return correct number of compiled functions");
-  }
-
-  for (size_t j = 0, end = nodes_and_viewers.size(); j < end; j++) {
+  // We will compile the fused nodes one by one, and fuse the subgraph if successful.
+  // If a compilation fails we undo the fusion and leave the original nodes available for other EPs to take
+  for (size_t j = 0, end = nodes_and_viewers.size(); j < end; ++j) {
     Node& node = nodes_and_viewers[j].fused_node;
+    std::vector<NodeComputeInfo> single_node_compute_func;
+    auto status = current_ep.Compile({nodes_and_viewers[j]}, single_node_compute_func);
+    if (!status.IsOK()) {
+      // There is compile error with the nodes_and_viewer[j], remove the fused_node and function from the graph
+      LOGS_DEFAULT(ERROR) << "EP: " << current_ep.Type() << " has Compile error: " << status.ErrorMessage();
+      graph.CancelFuseSubGraph(node);
+    } else {
+      ORT_RETURN_IF(single_node_compute_func.empty(), "single_node_compute_func should have 1 elements");
+      ORT_RETURN_IF_ERROR(func_mgr.AddFuncInfo(node.Name(), std::move(single_node_compute_func[0])));
 
-    ORT_RETURN_IF_ERROR(func_mgr.AddFuncInfo(node.Name(), std::move(node_compute_funcs[j])));
+      const auto& cur_capability = capabilities[j];
+      const IndexedSubGraph& indexed_sub_graph = *cur_capability->sub_graph;
+      const IndexedSubGraph::MetaDef& metadef = *indexed_sub_graph.GetMetaDef();
 
-    const auto& cur_capability = capabilities[j];
-    const IndexedSubGraph& indexed_sub_graph = *cur_capability->sub_graph;
-    const IndexedSubGraph::MetaDef& metadef = *indexed_sub_graph.GetMetaDef();
+      KernelDefBuilder builder;
+      BuildFusedKernelDef(builder, metadef, type);
+      auto kernel_def = builder.Build();
 
-    KernelDefBuilder builder;
-    BuildFusedKernelDef(builder, metadef, type);
-    auto kernel_def = builder.Build();
+      // save hash so SessionState can find the kernel. each kernel name should be unique
+      if (compiled_kernel_hashes.insert({metadef.name, kernel_def->GetHash()}).second == false) {
+        ORT_THROW("Existing entry in compiled kernel hashes for ", metadef.name,
+                  ". Execution Provider must generate unique names across the entire model.");
+      }
 
-    // save hash so SessionState can find the kernel. each kernel name should be unique
-    if (compiled_kernel_hashes.insert({metadef.name, kernel_def->GetHash()}).second == false) {
-      ORT_THROW("Existing entry in compiled kernel hashes for ", metadef.name,
-                ". Execution Provider must generate unique names across the entire model.");
+      ORT_RETURN_IF_ERROR(fused_kernel_registry.Register(
+          KernelCreateInfo(std::move(kernel_def), static_cast<KernelCreatePtrFn>(
+                                                      [](const OpKernelInfo& info) -> OpKernel* {
+                                                        return new FunctionKernel(info);
+                                                      }))));
+
+      // now that we're done compiling we can remove the original nodes from the Graph and wire in the new one
+      graph.FinalizeFuseSubGraph(indexed_sub_graph, node);
     }
-
-    ORT_RETURN_IF_ERROR(fused_kernel_registry.Register(
-        KernelCreateInfo(std::move(kernel_def), static_cast<KernelCreatePtrFn>(
-                                                    [](const OpKernelInfo& info) -> OpKernel* {
-                                                      return new FunctionKernel(info);
-                                                    }))));
-
-    // now that we're done compiling we can remove the original nodes from the Graph and wire in the new one
-    graph.FinalizeFuseSubGraph(indexed_sub_graph, node);
   }
 
   return Status::OK();

onnxruntime/core/graph/graph.cc

@@ -3392,6 +3392,31 @@ Node& Graph::BeginFuseSubGraph(const IndexedSubGraph& sub_graph, const std::stri
   return node;
 }
 
+void Graph::CancelFuseSubGraph(const Node& fused_node) {
+  auto node_idx = fused_node.Index();
+  if (!GetNode(node_idx))
+    return;
+
+  if (fused_node.NodeType() != Node::Type::Fused)
+    return;
+
+#if !defined(ORT_MINIMAL_BUILD)
+  // Remove the function body from function container
+  const auto* fused_node_func = fused_node.GetFunctionBody();
+  auto it = std::find_if(
+      function_container_.begin(), function_container_.end(),
+      [fused_node_func](const std::unique_ptr<onnxruntime::Function>& func) {
+        return func.get() == fused_node_func;
+      });
+  if (it != function_container_.end()) {
+    function_container_.erase(it);
+  }
+#endif
+
+  // Remove the fused_node
+  RemoveNode(node_idx);
+}
+
 void Graph::FinalizeFuseSubGraph(const IndexedSubGraph& sub_graph, Node& fused_node) {
   const auto* func_meta_def = sub_graph.GetMetaDef();
   ORT_ENFORCE(nullptr != func_meta_def);
@@ -3432,9 +3457,7 @@ void Graph::FinalizeFuseSubGraph(const IndexedSubGraph& sub_graph, Node& fused_n
         if (it != input_indexes.cend()) {
           AddEdge(producer_idx, new_node_idx, src_idx, it->second);
         }
-      } 
-      else
-      {
+      } else {
         int dst_implicit_input_idx = dst_idx - (int)node->InputDefs().size();
         ORT_ENFORCE(dst_implicit_input_idx < (int)node->ImplicitInputDefs().size());
         auto it = input_indexes.find(node->ImplicitInputDefs()[dst_implicit_input_idx]->Name());

onnxruntime/test/providers/internal_testing/internal_testing_tests.cc

@@ -254,5 +254,102 @@ TEST(InternalTestingEP, TestModelWithSubgraph) {
                             feeds);
 }
 
+// A custom InternalTestingEP extension
+// This is to testing execution fall back to CPU EP if Compile fails, for ORT format
+// This EP will take an additional compile_failure_ops
+// If in Compile() any nodes in the partition is also in compile_failure_ops
+// The Compile will fail
+class CompileFailureTestExecutionProvider : public InternalTestingExecutionProvider {
+ public:
+  CompileFailureTestExecutionProvider(const std::unordered_set<std::string>& supported_ops,
+                                      const std::unordered_set<std::string>& compile_failure_ops);
+  virtual ~CompileFailureTestExecutionProvider() = default;
+
+  Status Compile(const std::vector<FusedNodeAndGraph>& fused_nodes,
+                 std::vector<NodeComputeInfo>& node_compute_funcs) override;
+
+ private:
+  std::unordered_set<std::string> compile_failure_ops_;
+};
+
+CompileFailureTestExecutionProvider::CompileFailureTestExecutionProvider(
+    const std::unordered_set<std::string>& supported_ops,
+    const std::unordered_set<std::string>& compile_failure_ops)
+    : InternalTestingExecutionProvider(supported_ops),
+      compile_failure_ops_(compile_failure_ops) {}
+
+Status CompileFailureTestExecutionProvider::Compile(const std::vector<FusedNodeAndGraph>& fused_nodes,
+                                                    std::vector<NodeComputeInfo>& node_compute_funcs) {
+  for (const auto& fused_node_and_graph : fused_nodes) {
+    // If any nodes in this partition is also in compile_failure_ops_, the Compile will fail
+    const onnxruntime::GraphViewer& graph_viewer(fused_node_and_graph.filtered_graph);
+    for (const auto& node : graph_viewer.Nodes()) {
+      if (compile_failure_ops_.find(node.OpType()) != compile_failure_ops_.end()) {
+        return ORT_MAKE_STATUS(ONNXRUNTIME, FAIL,
+                               "CompileFailureTestExecutionProvider::Compile failed for node: ", node.Name());
+      }
+    }
+  }
+
+  return InternalTestingExecutionProvider::Compile(fused_nodes, node_compute_funcs);
+}
+
+TEST(InternalTestingEP, TestOrtModelWithCompileFailure) {
+  // In the test file, there are 2 Conv and 1 Gemm nodes, all disconnected
+  // So we should have 3 partitions be taken by InternalTestingExecutionProvider/CompileFailureTestExecutionProvider
+  // But CompileFailureTestExecutionProvider will fail the Compile for partition contains "Gemm" node
+  // This is to test the model initialization won't fail and Gemm node will not be replaced by the fused_node
+  const ORTCHAR_T* ort_model_path = ORT_TSTR("testdata/mnist.ort");
+
+  const std::unordered_set<std::string>& supported_ops{"Conv", "Gemm"};
+  const std::unordered_set<std::string>& compile_failure_ops{"Gemm"};
+
+  // Use InternalTestingExecutionProvider
+  // We should have 3 partitions taken by the EP
+  // 2 Conv and 1 Gemm
+  {
+    InferenceSessionWrapper session(SessionOptions(), GetEnvironment());
+    ASSERT_STATUS_OK(session.RegisterExecutionProvider(
+        onnxruntime::make_unique<InternalTestingExecutionProvider>(supported_ops)));
+    ASSERT_STATUS_OK(session.Load(ort_model_path));
+    ASSERT_STATUS_OK(session.Initialize());
+
+    int num_replaced_nodes = CountAndValidateAssignedNodes(
+        session.GetGraph(), supported_ops, session.GetSessionState().GetFuncMgr());
+
+    ASSERT_EQ(num_replaced_nodes, 3);
+  }
+
+  // Use CompileFailureTestExecutionProvider which will fail Compile on "Gemm"
+  // We should have 2 partitions taken by the EP
+  // 2 Conv
+  {
+    InferenceSessionWrapper session(SessionOptions(), GetEnvironment());
+    ASSERT_STATUS_OK(session.RegisterExecutionProvider(
+        onnxruntime::make_unique<CompileFailureTestExecutionProvider>(supported_ops, compile_failure_ops)));
+    ASSERT_STATUS_OK(session.Load(ort_model_path));
+    ASSERT_STATUS_OK(session.Initialize());
+
+    // 2 Conv nodes shoule be replaced with fused nodes
+    const auto& graph = session.GetGraph();
+    int num_replaced_nodes = CountAndValidateAssignedNodes(
+        session.GetGraph(), {"Conv"}, session.GetSessionState().GetFuncMgr());
+
+    ASSERT_EQ(num_replaced_nodes, 2);
+
+    // The Gemm node should still not have been replaced
+    int count_compile_failure_nodes = 0;
+    for (const auto& node : graph.Nodes()) {
+      if (compile_failure_ops.find(node.OpType()) != compile_failure_ops.end())
+        count_compile_failure_nodes++;
+    }
+    ASSERT_EQ(count_compile_failure_nodes, 1);
+
+    // Execute the session, since the last node is Gemm, and its input 0 is all 0s
+    // So the result should be the bias initializer of the Gemm node
+    ExecuteMnist(session, true /* enable_custom_ep */);
+  }
+}
+
 }  // namespace test
 }  // namespace onnxruntime