!5481 add mode black list checker

Merge pull request !5481 from zyli2020/master

mindspore/ccsrc/backend/session/gpu_session.cc

@@ -49,9 +49,10 @@ using AnfAlgo = mindspore::session::AnfRuntimeAlgorithm;
 
 void GPUSession::SelectKernel(const std::shared_ptr<KernelGraph> &kernel_graph) const {
   MS_EXCEPTION_IF_NULL(kernel_graph);
+  bool in_black_list = CheckInModeBlackList(kernel_graph);
   for (const auto &kernel_node : kernel_graph->execution_order()) {
     MS_EXCEPTION_IF_NULL(kernel_node);
-    device::gpu::SetKernelInfo(kernel_node);
+    device::gpu::SetKernelInfo(kernel_node, in_black_list);
   }
 }
 
@@ -75,7 +76,7 @@ void GPUSession::Optimize(const std::shared_ptr<KernelGraph> &kernel_graph) {
   pm->AddPass(std::make_shared<opt::ReplaceBNGradCastFusion>());
   pm->AddPass(std::make_shared<opt::ReplaceMomentumCastFusion>());
   pm->AddPass(std::make_shared<opt::ReplaceAddNFusion>());
-  if (context_ptr->execution_mode() != kPynativeMode) {
+  if (!CheckInModeBlackList(kernel_graph) && context_ptr->execution_mode() != kPynativeMode) {
     pm->AddPass(std::make_shared<opt::BatchNormReluFusion>());
     pm->AddPass(std::make_shared<opt::BatchNormReluGradFusion>());
     pm->AddPass(std::make_shared<opt::BatchNormAddReluFusion>());
@@ -192,6 +193,28 @@ void GPUSession::Execute(const std::shared_ptr<KernelGraph> &kernel_graph) const
   }
 }
 
+bool GPUSession::CheckInModeBlackList(const std::shared_ptr<KernelGraph> &kernel_graph) const {
+  auto kernels = kernel_graph->execution_order();
+  size_t conv_cnt = 0;
+  size_t bn_cnt = 0;
+  for (const auto &kernel : kernels) {
+    auto kernel_name = AnfAlgo::GetCNodeName(kernel);
+    if (kernel_name == prim::kPrimLayerNorm->name()) {
+      return true;
+    }
+    if (kernel_name == prim::kPrimConv2D->name()) {
+      conv_cnt++;
+    }
+    if (kernel_name == prim::kPrimFusedBatchNormEx->name()) {
+      bn_cnt++;
+    }
+  }
+  if (conv_cnt == kConv2dCount && bn_cnt == kFusedBatchNormCount) {
+    return true;
+  }
+  return false;
+}
+
 GraphId GPUSession::CompileGraph(const AnfNodePtrList &lst, const AnfNodePtrList &outputs) {
   // Construct graph, if successfully, graph_sum_ + 1
   auto graph_id = graph_sum_;

mindspore/ccsrc/backend/session/gpu_session.h

@@ -67,6 +67,8 @@ class GPUSession : public SessionBasic {
 
   void Execute(const std::shared_ptr<KernelGraph> &kernel_graph) const;
 
+  bool CheckInModeBlackList(const std::shared_ptr<KernelGraph> &kernel_graph) const;
+
 #ifdef ENABLE_DEBUGGER
   void Dump(const std::shared_ptr<KernelGraph> &kernel_graph) const;
 
@@ -80,6 +82,9 @@ class GPUSession : public SessionBasic {
 
   void PostLoadTensor(const std::shared_ptr<KernelGraph> &kernel_graph) const;
 #endif
+
+  static constexpr size_t kConv2dCount = 96;
+  static constexpr size_t kFusedBatchNormCount = 94;
 };
 using GPUSessionPtr = std::shared_ptr<GPUSession>;
 MS_REG_SESSION(kGPUDevice, GPUSession);

mindspore/ccsrc/runtime/device/gpu/kernel_info_setter.cc

@@ -223,7 +223,7 @@ void UpdateKernelFormatInfo(const CNodePtr &kernel_node, const std::vector<TypeI
 }
 }  // namespace
 
-void SetKernelInfo(const CNodePtr &kernel_node) {
+void SetKernelInfo(const CNodePtr &kernel_node, bool in_black_list) {
   std::vector<std::string> inputs_format;
   std::vector<TypeId> inputs_type;
   for (size_t input_index = 0; input_index < AnfAlgo::GetInputTensorNum(kernel_node); ++input_index) {
@@ -237,7 +237,7 @@ void SetKernelInfo(const CNodePtr &kernel_node) {
     outputs_type.push_back(AnfAlgo::GetOutputInferDataType(kernel_node, output_index));
   }
   std::string origin_data_format = kOpFormat_DEFAULT;
-  if (IsNeedProcessFormatInfo(kernel_node, inputs_type)) {
+  if (!in_black_list && IsNeedProcessFormatInfo(kernel_node, inputs_type)) {
     UpdateKernelFormatInfo(kernel_node, inputs_type, &inputs_format, &outputs_format, &origin_data_format);
   }
   std::shared_ptr<KernelBuildInfo::KernelBuildInfoBuilder> builder =

mindspore/ccsrc/runtime/device/gpu/kernel_info_setter.h

@@ -53,7 +53,7 @@ static std::map<std::string, std::pair<std::vector<size_t>, std::vector<size_t>>
   {prim::kPrimAddN->name(), {{}, {0}}},
 };
 
-void SetKernelInfo(const CNodePtr &apply_kernel_ptr);
+void SetKernelInfo(const CNodePtr &kernel_node, bool in_black_list = false);
 
 class KernelAttr {
  public: