optimize opencl runtime

mindspore/lite/src/lite_session.cc

@@ -26,9 +26,6 @@
 #include "src/common/utils.h"
 #include "src/common/graph_util.h"
 #include "src/kernel_registry.h"
-#if SUPPORT_GPU
-#include "src/runtime/opencl/opencl_runtime.h"
-#endif
 
 namespace mindspore {
 namespace lite {
@@ -343,7 +340,7 @@ int LiteSession::Init(Context *context) {
   }
 #if SUPPORT_GPU
   if (context_->device_type_ == DT_GPU) {
-    auto opencl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+    auto opencl_runtime = ocl_runtime_wrap_.GetInstance();
     opencl_runtime->SetFp16Enable(context_->float16_priority);
     if (opencl_runtime->Init() != RET_OK) {
       context_->device_type_ = DT_CPU;
@@ -394,11 +391,6 @@ LiteSession::~LiteSession() {
   for (auto *kernel : kernels_) {
     delete kernel;
   }
-#if SUPPORT_GPU
-  if (context_->device_type_ == DT_GPU) {
-    lite::opencl::OpenCLRuntime::DeleteInstance();
-  }
-#endif
   delete this->context_;
   delete this->executor;
   this->executor = nullptr;

mindspore/lite/src/lite_session.h

@@ -30,6 +30,9 @@
 #include "schema/model_generated.h"
 #include "src/executor.h"
 #include "src/tensor.h"
+#if SUPPORT_GPU
+#include "src/runtime/opencl/opencl_runtime.h"
+#endif
 
 namespace mindspore {
 namespace lite {
@@ -108,6 +111,9 @@ class LiteSession : public session::LiteSession {
   std::unordered_map<std::string, mindspore::tensor::MSTensor *> output_tensor_map_;
   Executor *executor = nullptr;
   std::atomic<bool> is_running_ = false;
+#if SUPPORT_GPU
+  opencl::OpenCLRuntimeWrapper ocl_runtime_wrap_;
+#endif
 };
 }  // namespace lite
 }  // namespace mindspore

mindspore/lite/src/runtime/kernel/opencl/kernel/gather.cc

@@ -54,7 +54,7 @@ int GatherOpenCLKernel::Init() {
   auto indices_tensor = in_tensors_.at(1);
   int indices_num = indices_tensor->ElementsNum();
   bool isIndicesInt32 = indices_tensor->data_type() == kNumberTypeInt32;
-  auto allocator = lite::opencl::OpenCLRuntime::GetInstance()->GetAllocator();
+  auto allocator = ocl_runtime_->GetAllocator();
   if (!isIndicesInt32) {
     indices_data_ = reinterpret_cast<int32_t *>(allocator->Malloc(sizeof(int32_t) * indices_num));
     if (indices_data_ == nullptr) {

mindspore/lite/src/runtime/kernel/opencl/opencl_kernel.h

@@ -38,15 +38,10 @@ class OpenCLKernel : public LiteKernel {
   explicit OpenCLKernel(OpParameter *parameter, const std::vector<lite::Tensor *> &inputs,
                         const std::vector<lite::Tensor *> &outputs)
       : LiteKernel(parameter, inputs, outputs, nullptr, nullptr) {
-    ocl_runtime_ = lite::opencl::OpenCLRuntime::GetInstance();
+    ocl_runtime_ = ocl_runtime_wrap_.GetInstance();
   }
 
-  ~OpenCLKernel() {
-    if (ocl_runtime_ != nullptr) {
-      lite::opencl::OpenCLRuntime::DeleteInstance();
-      ocl_runtime_ = nullptr;
-    }
-  }
+  ~OpenCLKernel() {}
 
   virtual int Init() { return RET_ERROR; }
   virtual int Prepare() { return RET_ERROR; }
@@ -69,7 +64,8 @@ class OpenCLKernel : public LiteKernel {
   schema::Format in_ori_format_{schema::Format::Format_NHWC};
   schema::Format out_ori_format_{schema::Format::Format_NHWC4};
   schema::Format op_format_{schema::Format::Format_NHWC4};
-  lite::opencl::OpenCLRuntime *ocl_runtime_{nullptr};
+  lite::opencl::OpenCLRuntimeWrapper ocl_runtime_wrap_;
+  lite::opencl::OpenCLRuntime *ocl_runtime_;
 };
 }  // namespace mindspore::kernel
 

mindspore/lite/src/runtime/kernel/opencl/subgraph_opencl_kernel.cc

@@ -17,7 +17,6 @@
 #include "src/runtime/kernel/opencl/subgraph_opencl_kernel.h"
 #include <set>
 #include "src/runtime/opencl/opencl_executor.h"
-#include "src/runtime/opencl/opencl_runtime.h"
 #include "src/runtime/kernel/opencl/utils.h"
 #include "include/errorcode.h"
 #include "src/common/utils.h"
@@ -161,7 +160,6 @@ int SubGraphOpenCLKernel::GenToFormatOp(const std::vector<lite::Tensor *> &in_te
 }
 
 int SubGraphOpenCLKernel::Init() {
-  ocl_runtime_ = lite::opencl::OpenCLRuntime::GetInstance();
   allocator_ = ocl_runtime_->GetAllocator();
   MS_LOG(DEBUG) << "input num=" << in_tensors_.size() << ", output num=" << out_tensors_.size();
   for (const auto tensor : in_tensors_) {
@@ -308,10 +306,6 @@ int SubGraphOpenCLKernel::UnInit() {
   nodes_.clear();
   in_convert_ops_.clear();
   out_convert_ops_.clear();
-  if (ocl_runtime_ != nullptr) {
-    lite::opencl::OpenCLRuntime::DeleteInstance();
-    ocl_runtime_ = nullptr;
-  }
   return RET_OK;
 }
 

mindspore/lite/src/runtime/kernel/opencl/subgraph_opencl_kernel.h

@@ -36,7 +36,9 @@ class SubGraphOpenCLKernel : public SubGraphKernel {
                                 const std::vector<kernel::LiteKernel *> outKernels,
                                 const std::vector<kernel::LiteKernel *> nodes, const lite::InnerContext *ctx = nullptr,
                                 const mindspore::lite::PrimitiveC *primitive = nullptr)
-      : SubGraphKernel(inputs, outputs, inKernels, outKernels, nodes, ctx, primitive) {}
+      : SubGraphKernel(inputs, outputs, inKernels, outKernels, nodes, ctx, primitive) {
+    ocl_runtime_ = ocl_runtime_wrap_.GetInstance();
+  }
   ~SubGraphOpenCLKernel() override;
 
   int Init() override;
@@ -64,6 +66,7 @@ class SubGraphOpenCLKernel : public SubGraphKernel {
   std::vector<OpenCLToFormatParameter *> out_parameters_;
   std::vector<LiteKernel *> in_convert_ops_;
   std::vector<LiteKernel *> out_convert_ops_;
+  lite::opencl::OpenCLRuntimeWrapper ocl_runtime_wrap_;
   lite::opencl::OpenCLRuntime *ocl_runtime_{nullptr};
 };
 }  // namespace mindspore::kernel

mindspore/lite/src/runtime/kernel/opencl/utils.cc

@@ -15,9 +15,11 @@
  */
 
 #include "src/runtime/kernel/opencl/utils.h"
+#include <fstream>
 #include <algorithm>
 #include <vector>
 #include "src/kernel_registry.h"
+#include "src/runtime/opencl/opencl_runtime.h"
 
 using mindspore::lite::KernelRegistrar;
 
@@ -221,4 +223,64 @@ std::string CLErrorCode(cl_int error_code) {
       return "Unknown OpenCL error code";
   }
 }
+
+void Write2File(void *mem, const std::string &file_name, int size) {
+  std::fstream os;
+  os.open(file_name, std::ios::out | std::ios::binary);
+  os.write(static_cast<char *>(mem), size);
+  os.close();
+}
+
+void PrintTensor(lite::Tensor *tensor, int num, const std::string &out_file) {
+  if (tensor->data_c() == nullptr) {
+    return;
+  }
+  auto runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
+  runtime->SyncCommandQueue();
+
+  auto allocator = runtime->GetAllocator();
+  auto origin_data = tensor->data_c();
+  allocator->MapBuffer(origin_data, CL_MAP_READ, nullptr, true);
+  tensor->SetData(origin_data);
+
+  auto Height = tensor->shape().size() == 4 ? tensor->Height() : 1;
+  auto Width = tensor->shape().size() == 4 ? tensor->Width() : 1;
+  auto SLICES = UP_DIV(tensor->Channel(), C4NUM);
+  auto alignment = runtime->GetImagePitchAlignment();
+  auto dtype_size = tensor->data_type() == kNumberTypeFloat16 ? sizeof(cl_half4) : sizeof(cl_float4);
+  auto row_pitch = (Width * SLICES + alignment - 1) / alignment * alignment * dtype_size;
+  auto row_size = Width * SLICES * dtype_size;
+  std::cout << "tensor->GetFormat() =" << tensor->GetFormat() << "\n";
+  std::cout << "Height              =" << Height << "\n";
+  std::cout << "Width               =" << Width << "\n";
+  std::cout << "SLICES              =" << SLICES << "\n";
+  std::cout << "image_alignment     =" << alignment << "\n";
+  std::cout << "dtype_size          =" << dtype_size << "\n";
+  std::cout << "row_pitch           =" << row_pitch << "\n";
+  std::cout << "row_size            =" << row_size << "\n";
+  std::cout << "tensor->Size()      =" << tensor->Size() << "\n";
+  std::vector<char> data(tensor->Size());
+  for (int i = 0; i < Height; ++i) {
+    memcpy(static_cast<char *>(data.data()) + i * row_size, static_cast<char *>(origin_data) + i * row_pitch, row_size);
+  }
+
+  std::cout << "shape=(";
+  for (auto x : tensor->shape()) {
+    printf("%3d,", x);
+  }
+  printf("): ");
+
+  for (size_t i = 0; i < num && i < tensor->ElementsNum(); ++i) {
+    if (tensor->data_type() == kNumberTypeFloat16)
+      printf("%zu %6.3f | ", i, (reinterpret_cast<float16_t *>(data.data()))[i]);
+    else
+      printf("%zu %6.3f | ", i, (reinterpret_cast<float *>(data.data()))[i]);
+  }
+  printf("\n");
+
+  if (!out_file.empty()) {
+    Write2File(data.data(), out_file, tensor->Size());
+  }
+  allocator->UnmapBuffer(origin_data);
+}
 }  // namespace mindspore::kernel

mindspore/lite/src/runtime/kernel/opencl/utils.h

@@ -44,6 +44,10 @@ std::vector<size_t> GetCommonLocalSize(const std::vector<size_t> &global, int ma
 
 std::string CLErrorCode(cl_int error_code);
 
+void Write2File(void *mem, const std::string &file_name, int size);
+
+void PrintTensor(lite::Tensor *tensor, int num = 10, const std::string &out_file = "");
+
 template <class T1, class T2>
 void PackNCHWToNC4HW4(void *src, void *dst, int batch, int plane, int channel, const std::function<T2(T1)> &to_dtype) {
   int c4 = UP_DIV(channel, C4NUM);

mindspore/lite/src/runtime/opencl/opencl_executor.h

@@ -27,11 +27,7 @@
 namespace mindspore::lite::opencl {
 class OpenCLExecutor : Executor {
  public:
-  OpenCLExecutor() : Executor() {
-    auto ocl_runtime = OpenCLRuntime::GetInstance();
-    allocator_ = ocl_runtime->GetAllocator();
-    OpenCLRuntime::DeleteInstance();
-  }
+  OpenCLExecutor() : Executor() { allocator_ = ocl_runtime.GetInstance()->GetAllocator(); }
 
   int Prepare(const std::vector<kernel::LiteKernel *> &kernels);
 
@@ -42,6 +38,7 @@ class OpenCLExecutor : Executor {
  protected:
   InnerContext *context = nullptr;
   OpenCLAllocator *allocator_;
+  OpenCLRuntimeWrapper ocl_runtime;
 };
 }  // namespace mindspore::lite::opencl
 #endif

mindspore/lite/src/runtime/opencl/opencl_runtime.cc

@@ -393,11 +393,16 @@ int OpenCLRuntime::RunKernel(const cl::Kernel &kernel, const std::vector<size_t>
   cl::Event event;
   cl_int ret = CL_SUCCESS;
   ret = command_queue->enqueueNDRangeKernel(kernel, cl::NullRange, global_range, local_range, nullptr, &event);
-
   if (ret != CL_SUCCESS) {
     MS_LOG(ERROR) << "Kernel execute failed:" << CLErrorCode(ret);
     return RET_ERROR;
   }
+  static int cnt = 0;
+  const int flush_period = 10;
+  if (cnt % flush_period == 0) {
+    command_queue->flush();
+  }
+  cnt++;
   MS_LOG(DEBUG) << "RunKernel success!";
 #if MS_OPENCL_PROFILE
   event.wait();

mindspore/lite/src/runtime/opencl/opencl_runtime.h

@@ -37,11 +37,10 @@ struct GpuInfo {
   int model_num = 0;
   float opencl_version = 0;
 };
-
+class OpenCLRuntimeWrapper;
 class OpenCLRuntime {
  public:
-  static OpenCLRuntime *GetInstance();
-  static void DeleteInstance();
+  friend OpenCLRuntimeWrapper;
 
   ~OpenCLRuntime();
   OpenCLRuntime(const OpenCLRuntime &) = delete;
@@ -138,6 +137,8 @@ class OpenCLRuntime {
   int GetKernelMaxWorkGroupSize(cl_kernel kernel, cl_device_id device_id);
 
  private:
+  static OpenCLRuntime *GetInstance();
+  static void DeleteInstance();
   OpenCLRuntime();
   GpuInfo ParseGpuInfo(std::string device_name, std::string device_version);
 
@@ -169,5 +170,16 @@ class OpenCLRuntime {
   void *handle_{nullptr};
 };
 
+class OpenCLRuntimeWrapper {
+ public:
+  OpenCLRuntimeWrapper() { ocl_runtime_ = OpenCLRuntime::GetInstance(); }
+  ~OpenCLRuntimeWrapper() { OpenCLRuntime::DeleteInstance(); }
+  explicit OpenCLRuntimeWrapper(const OpenCLRuntime &) = delete;
+  OpenCLRuntimeWrapper &operator=(const OpenCLRuntime &) = delete;
+  OpenCLRuntime *GetInstance() { return ocl_runtime_; }
+
+ private:
+  OpenCLRuntime *ocl_runtime_{nullptr};
+};
 }  // namespace mindspore::lite::opencl
 #endif  // MINDSPORE_LITE_SRC_OPENCL_RUNTIME_H_

mindspore/lite/test/ut/src/runtime/kernel/opencl/activation_tests.cc

@@ -82,7 +82,7 @@ TEST_F(TestActivationOpenCL, ReluFp_dim4) {
   std::string in_file = "/data/local/tmp/in_data.bin";
   std::string out_file = "/data/local/tmp/relu.bin";
   MS_LOG(INFO) << "Relu Begin test!";
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   auto allocator = ocl_runtime->GetAllocator();
   auto data_type = kNumberTypeFloat16;
@@ -184,14 +184,13 @@ TEST_F(TestActivationOpenCL, ReluFp_dim4) {
   delete input_tensor;
   delete output_tensor;
   delete sub_graph;
-  lite::opencl::OpenCLRuntime::DeleteInstance();
 }
 
 TEST_F(TestActivationOpenCL, Relu6Fp_dim4) {
   std::string in_file = "/data/local/tmp/in_data.bin";
   std::string out_file = "/data/local/tmp/relu6.bin";
   MS_LOG(INFO) << "Relu6 Begin test!";
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   auto data_type = kNumberTypeFloat16;
   ocl_runtime->SetFp16Enable(data_type == kNumberTypeFloat16);
   bool enable_fp16 = ocl_runtime->GetFp16Enable();
@@ -296,14 +295,13 @@ TEST_F(TestActivationOpenCL, Relu6Fp_dim4) {
   delete input_tensor;
   delete output_tensor;
   delete sub_graph;
-  lite::opencl::OpenCLRuntime::DeleteInstance();
 }
 
 TEST_F(TestActivationOpenCL, SigmoidFp_dim4) {
   std::string in_file = "/data/local/tmp/in_data.bin";
   std::string out_file = "/data/local/tmp/sigmoid.bin";
   MS_LOG(INFO) << "Sigmoid Begin test!";
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   auto data_type = kNumberTypeFloat32;
   ocl_runtime->SetFp16Enable(data_type == kNumberTypeFloat16);
@@ -408,14 +406,13 @@ TEST_F(TestActivationOpenCL, SigmoidFp_dim4) {
   delete input_tensor;
   delete output_tensor;
   delete sub_graph;
-  lite::opencl::OpenCLRuntime::DeleteInstance();
 }
 
 TEST_F(TestActivationOpenCL, LeakyReluFp_dim4) {
   std::string in_file = "/data/local/tmp/in_data.bin";
   std::string out_file = "/data/local/tmp/leaky_relu.bin";
   MS_LOG(INFO) << "Leaky relu Begin test!";
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   auto data_type = kNumberTypeFloat16;  // need modify
   ocl_runtime->SetFp16Enable(data_type == kNumberTypeFloat16);
@@ -519,14 +516,13 @@ TEST_F(TestActivationOpenCL, LeakyReluFp_dim4) {
   delete param;
   delete input_tensor;
   delete output_tensor;
-  lite::opencl::OpenCLRuntime::DeleteInstance();
 }
 
 TEST_F(TestActivationOpenCLTanh, TanhFp_dim4) {
   std::string in_file = "/data/local/tmp/test_data/in_tanhfp16.bin";
   std::string out_file = "/data/local/tmp/test_data/out_tanhfp16.bin";
   MS_LOG(INFO) << "Tanh Begin test!";
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   auto data_type = kNumberTypeFloat16;
   ocl_runtime->SetFp16Enable(data_type == kNumberTypeFloat16);
@@ -627,7 +623,6 @@ TEST_F(TestActivationOpenCLTanh, TanhFp_dim4) {
     printf_tensor<float>("Tanh:FP32--output data---", outputs[0]);
     CompareRes<float>(output_tensor, out_file);
   }
-  lite::opencl::OpenCLRuntime::DeleteInstance();
   input_tensor->SetData(nullptr);
   delete input_tensor;
   output_tensor->SetData(nullptr);

mindspore/lite/test/ut/src/runtime/kernel/opencl/arithmetic_self_tests.cc

@@ -43,7 +43,7 @@ void CompareOutputData1(T *input_data1, T *output_data, T *correct_data, int siz
 
 TEST_F(TestArithmeticSelfOpenCLfp16, ArithmeticSelfOpenCLFp16) {
   MS_LOG(INFO) << " begin test ";
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->SetFp16Enable(true);
   ocl_runtime->Init();
   auto allocator = ocl_runtime->GetAllocator();
@@ -125,7 +125,6 @@ TEST_F(TestArithmeticSelfOpenCLfp16, ArithmeticSelfOpenCLFp16) {
   sub_graph->Run();
   auto *output_data_gpu = reinterpret_cast<float16_t *>(output_tensor->data_c());
   CompareOutputData1(input_data1, output_data_gpu, correctOutput, output_tensor->ElementsNum(), 0.000001);
-  lite::opencl::OpenCLRuntime::DeleteInstance();
   for (auto tensor : inputs) {
     tensor->SetData(nullptr);
     delete tensor;
@@ -139,7 +138,7 @@ TEST_F(TestArithmeticSelfOpenCLfp16, ArithmeticSelfOpenCLFp16) {
 
 TEST_F(TestArithmeticSelfOpenCLCI, ArithmeticSelfRound) {
   MS_LOG(INFO) << " begin test ";
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   auto allocator = ocl_runtime->GetAllocator();
   float input_data1[] = {0.75f, 0.06f, 0.74f, 0.30f, 0.9f, 0.59f, 0.03f, 0.37f,
@@ -216,7 +215,6 @@ TEST_F(TestArithmeticSelfOpenCLCI, ArithmeticSelfRound) {
   sub_graph->Run();
   auto *output_data_gpu = reinterpret_cast<float *>(output_tensor->data_c());
   CompareOutputData1(input_data1, output_data_gpu, correctOutput, output_tensor->ElementsNum(), 0.000001);
-  lite::opencl::OpenCLRuntime::DeleteInstance();
   for (auto tensor : inputs) {
     tensor->SetData(nullptr);
     delete tensor;

mindspore/lite/test/ut/src/runtime/kernel/opencl/arithmetic_tests.cc

@@ -68,7 +68,7 @@ static void LogData(void *data, const int size, const std::string prefix) {
 template <class T>
 static void TestCase(const std::vector<int> &shape_a, const std::vector<int> &shape_b) {
   bool is_log_data = false;
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   auto allocator = ocl_runtime->GetAllocator();
 
   bool is_bias_add = shape_b.empty();
@@ -212,7 +212,6 @@ static void TestCase(const std::vector<int> &shape_a, const std::vector<int> &sh
   for (auto tensor : outputs) {
     delete tensor;
   }
-  lite::opencl::OpenCLRuntime::DeleteInstance();
 }
 
 class TestArithmeticOpenCL : public mindspore::CommonTest {

mindspore/lite/test/ut/src/runtime/kernel/opencl/avg_pooling_tests.cc

@@ -53,7 +53,7 @@ void InitAvgPoolingParam(PoolingParameter *param) {
 }
 
 void RunTestCaseAvgPooling(const std::vector<int> &shape, void *input_data, void *output_data, bool enable_fp16) {
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   size_t dtype_size = enable_fp16 ? sizeof(float16_t) : sizeof(float);
   ocl_runtime->SetFp16Enable(enable_fp16);
@@ -125,7 +125,6 @@ void RunTestCaseAvgPooling(const std::vector<int> &shape, void *input_data, void
   }
 
   MS_LOG(INFO) << "Test AvgPool2d passed";
-  lite::opencl::OpenCLRuntime::DeleteInstance();
 }
 
 TEST_F(TestAvgPoolingOpenCL, AvgPoolingFp32) {

mindspore/lite/test/ut/src/runtime/kernel/opencl/batchnorm_tests.cc

@@ -38,7 +38,7 @@ class TestBatchnormOpenCLCI : public mindspore::CommonTest {
 
 TEST_F(TestBatchnormOpenCLCI, Batchnormfp32CI) {
   MS_LOG(INFO) << " begin test ";
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   auto allocator = ocl_runtime->GetAllocator();
 
@@ -142,7 +142,6 @@ TEST_F(TestBatchnormOpenCLCI, Batchnormfp32CI) {
 
   auto *output_data_gpu = reinterpret_cast<float *>(output_tensor->data_c());
   CompareOutputData(output_data_gpu, correct_data, output_tensor->ElementsNum(), 0.0001);
-  lite::opencl::OpenCLRuntime::DeleteInstance();
   for (auto tensor : inputs) {
     tensor->SetData(nullptr);
     delete tensor;
@@ -156,7 +155,7 @@ TEST_F(TestBatchnormOpenCLCI, Batchnormfp32CI) {
 
 TEST_F(TestBatchnormOpenCLfp16, Batchnormfp16input_dim4) {
   MS_LOG(INFO) << "begin test";
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->SetFp16Enable(true);
   ocl_runtime->Init();
   auto allocator = ocl_runtime->GetAllocator();
@@ -262,7 +261,6 @@ TEST_F(TestBatchnormOpenCLfp16, Batchnormfp16input_dim4) {
 
   auto *output_data_gpu = reinterpret_cast<float16_t *>(output_tensor->data_c());
   CompareOutputData(output_data_gpu, correct_data, output_tensor->ElementsNum(), 0.01);
-  lite::opencl::OpenCLRuntime::DeleteInstance();
   for (auto tensor : inputs) {
     tensor->SetData(nullptr);
     delete tensor;
@@ -276,7 +274,7 @@ TEST_F(TestBatchnormOpenCLfp16, Batchnormfp16input_dim4) {
 
 TEST_F(TestBatchnormOpenCLfp32, Batchnormfp32input_dim4) {
   MS_LOG(INFO) << " begin test ";
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   auto allocator = ocl_runtime->GetAllocator();
 
@@ -381,7 +379,6 @@ TEST_F(TestBatchnormOpenCLfp32, Batchnormfp32input_dim4) {
 
   auto *output_data_gpu = reinterpret_cast<float *>(output_tensor->data_c());
   CompareOutputData(output_data_gpu, correct_data, output_tensor->ElementsNum(), 0.0001);
-  lite::opencl::OpenCLRuntime::DeleteInstance();
   for (auto tensor : inputs) {
     tensor->SetData(nullptr);
     delete tensor;

mindspore/lite/test/ut/src/runtime/kernel/opencl/biasadd_tests.cc

@@ -75,7 +75,7 @@ TEST_F(TestBiasAddOpenCL, BiasAddFp32_dim4) {
   std::string weight_file = "/data/local/tmp/weight_data.bin";
   std::string standard_answer_file = "/data/local/tmp/biasadd.bin";
   MS_LOG(INFO) << "BiasAdd Begin test:";
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   auto data_type = kNumberTypeFloat16;  // need modify
   ocl_runtime->SetFp16Enable(data_type == kNumberTypeFloat16);
@@ -200,6 +200,5 @@ TEST_F(TestBiasAddOpenCL, BiasAddFp32_dim4) {
   delete output_tensor;
   delete sub_graph;
   delete param;
-  lite::opencl::OpenCLRuntime::DeleteInstance();
 }
 }  // namespace mindspore

mindspore/lite/test/ut/src/runtime/kernel/opencl/cast_tests.cc

@@ -38,7 +38,7 @@ void CompareOutputData1(T *output_data, T *correct_data, int size, float err_bou
 
 TEST_F(TestCastSelfOpenCL, Castfp32tofp16) {
   MS_LOG(INFO) << " begin test ";
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   auto allocator = ocl_runtime->GetAllocator();
 
@@ -113,7 +113,6 @@ TEST_F(TestCastSelfOpenCL, Castfp32tofp16) {
   sub_graph->Run();
   auto *output_data_gpu = reinterpret_cast<float16_t *>(output_tensor->data_c());
   CompareOutputData1(output_data_gpu, correctOutput, output_tensor->ElementsNum(), 0.000001);
-  lite::opencl::OpenCLRuntime::DeleteInstance();
   for (auto tensor : inputs) {
     tensor->SetData(nullptr);
     delete tensor;
@@ -127,7 +126,7 @@ TEST_F(TestCastSelfOpenCL, Castfp32tofp16) {
 
 TEST_F(TestCastSelfOpenCL, Castfp16tofp32) {
   MS_LOG(INFO) << " begin test ";
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   auto allocator = ocl_runtime->GetAllocator();
 
@@ -201,7 +200,6 @@ TEST_F(TestCastSelfOpenCL, Castfp16tofp32) {
   sub_graph->Run();
   auto *output_data_gpu = reinterpret_cast<float *>(output_tensor->data_c());
   CompareOutputData1(output_data_gpu, correctOutput, output_tensor->ElementsNum(), 0.000001);
-  lite::opencl::OpenCLRuntime::DeleteInstance();
   for (auto tensor : inputs) {
     tensor->SetData(nullptr);
     delete tensor;

mindspore/lite/test/ut/src/runtime/kernel/opencl/concat_tests.cc

@@ -47,7 +47,7 @@ void CompareOutputData1(T *output_data, T *correct_data, int size, float err_bou
 
 TEST_F(TestConcatOpenCLCI, ConcatFp32_2inputforCI) {
   MS_LOG(INFO) << " begin test ";
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   auto allocator = ocl_runtime->GetAllocator();
 
@@ -134,7 +134,6 @@ TEST_F(TestConcatOpenCLCI, ConcatFp32_2inputforCI) {
   sub_graph->Run();
   auto *output_data_gpu = reinterpret_cast<float *>(output_tensor->data_c());
   CompareOutputData1(output_data_gpu, correctOutput, output_tensor->ElementsNum(), 0.00001);
-  lite::opencl::OpenCLRuntime::DeleteInstance();
   for (auto tensor : inputs) {
     tensor->SetData(nullptr);
     delete tensor;
@@ -148,7 +147,7 @@ TEST_F(TestConcatOpenCLCI, ConcatFp32_2inputforCI) {
 
 TEST_F(TestConcatOpenCLfp16, ConcatFp16_2input_dim4_axis1) {
   MS_LOG(INFO) << " begin test ";
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->SetFp16Enable(true);
   ocl_runtime->Init();
   auto allocator = ocl_runtime->GetAllocator();
@@ -264,7 +263,6 @@ TEST_F(TestConcatOpenCLfp16, ConcatFp16_2input_dim4_axis1) {
   sub_graph->Run();
   auto *output_data_gpu = reinterpret_cast<float16_t *>(output_tensor->data_c());
   CompareOutputData1(output_data_gpu, correctOutput, output_tensor->ElementsNum(), 0.000001);
-  lite::opencl::OpenCLRuntime::DeleteInstance();
   for (auto tensor : inputs) {
     tensor->SetData(nullptr);
     delete tensor;
@@ -278,7 +276,7 @@ TEST_F(TestConcatOpenCLfp16, ConcatFp16_2input_dim4_axis1) {
 
 TEST_F(TestConcatOpenCLfp32, ConcatFp32_2input_dim4_axis3) {
   MS_LOG(INFO) << " begin test ";
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   auto allocator = ocl_runtime->GetAllocator();
 
@@ -385,7 +383,6 @@ TEST_F(TestConcatOpenCLfp32, ConcatFp32_2input_dim4_axis3) {
   sub_graph->Run();
   auto *output_data_gpu = reinterpret_cast<float *>(output_tensor->data_c());
   CompareOutputData1(output_data_gpu, correctOutput, output_tensor->ElementsNum(), 0.00001);
-  lite::opencl::OpenCLRuntime::DeleteInstance();
   for (auto tensor : inputs) {
     tensor->SetData(nullptr);
     delete tensor;

mindspore/lite/test/ut/src/runtime/kernel/opencl/conv2d_transpose_tests.cc

@@ -32,7 +32,7 @@ class TestConv2dTransposeOpenCL : public mindspore::CommonTest {
 
 void RunTestCaseConv2dTranspose(const std::vector<int> &shape, void *input_data, void *weight_data, void *bias_data,
                                 void *output_data, bool enable_fp16) {
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   size_t dtype_size = enable_fp16 ? sizeof(float16_t) : sizeof(float);
   ocl_runtime->SetFp16Enable(enable_fp16);
@@ -134,7 +134,6 @@ void RunTestCaseConv2dTranspose(const std::vector<int> &shape, void *input_data,
   for (auto t : outputs) {
     t->SetData(nullptr);
   }
-  lite::opencl::OpenCLRuntime::DeleteInstance();
 }
 
 TEST_F(TestConv2dTransposeOpenCL, Conv2dTransposeFp32) {

mindspore/lite/test/ut/src/runtime/kernel/opencl/convolution_tests.cc

@@ -157,7 +157,7 @@ void TEST_MAIN(const std::string &attr, Format input_format, Format output_forma
          &param->dilation_h_, &param->dilation_w_);
 
   MS_LOG(DEBUG) << "initialize OpenCLRuntime and OpenCLAllocator";
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   ocl_runtime->SetFp16Enable(data_type == kNumberTypeFloat16);
   auto allocator = ocl_runtime->GetAllocator();
@@ -201,7 +201,6 @@ void TEST_MAIN(const std::string &attr, Format input_format, Format output_forma
   input.SetData(nullptr);
   output.SetData(nullptr);
   delete sub_graph;
-  lite::opencl::OpenCLRuntime::DeleteInstance();
 }
 
 void TEST_MAIN(const std::string &attr, Format input_format, Format output_format, const TypeId data_type,

mindspore/lite/test/ut/src/runtime/kernel/opencl/depthwise_conv2d_tests.cc

@@ -33,7 +33,7 @@ class TestConvolutionDwOpenCL : public mindspore::CommonTest {
 template <class T1, class T2>
 void DepthWiseTestMain(ConvParameter *conv_param, T2 *input_data, T1 *weight_data, T2 *gnd_data, schema::Format format,
                        TypeId dtype = kNumberTypeFloat32, bool is_compare = true, T2 err_max = 1e-5) {
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   auto allocator = ocl_runtime->GetAllocator();
   if (dtype == kNumberTypeFloat16) {
@@ -167,7 +167,6 @@ void DepthWiseTestMain(ConvParameter *conv_param, T2 *input_data, T1 *weight_dat
   inputs[1]->SetData(nullptr);
   inputs[2]->SetData(nullptr);
   delete[] packed_input;
-  lite::opencl::OpenCLRuntime::DeleteInstance();
   inputs[0]->SetData(nullptr);
   outputs[0]->SetData(nullptr);
   return;

mindspore/lite/test/ut/src/runtime/kernel/opencl/gather_tests.cc

@@ -32,7 +32,7 @@ void test_main_gather(void *input_data, void *correct_data, const std::vector<in
                       const std::vector<int> &indices, GatherParameter *param, TypeId data_type,
                       schema::Format format) {
   MS_LOG(INFO) << " begin test ";
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   auto allocator = ocl_runtime->GetAllocator();
 

mindspore/lite/test/ut/src/runtime/kernel/opencl/matmul_tests.cc

@@ -31,7 +31,7 @@ class TestMatMulOpenCL : public mindspore::CommonTest {
 
 void RunTestCaseMatMul(const std::vector<int> &shape, void *input_data, void *weight_data, void *output_data,
                        bool enable_fp16, int dims) {
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   size_t dtype_size = enable_fp16 ? sizeof(float16_t) : sizeof(float);
   ocl_runtime->SetFp16Enable(enable_fp16);
@@ -123,7 +123,6 @@ void RunTestCaseMatMul(const std::vector<int> &shape, void *input_data, void *we
     t->SetData(nullptr);
   }
   MS_LOG(INFO) << "TestMatMul passed";
-  lite::opencl::OpenCLRuntime::DeleteInstance();
 }
 
 TEST_F(TestMatMulOpenCL, MatMul2DFp32) {

mindspore/lite/test/ut/src/runtime/kernel/opencl/max_pooling_tests.cc

@@ -53,7 +53,7 @@ void InitMaxPoolingParam(PoolingParameter *param) {
 }
 
 void RunTestCaseMaxPooling(const std::vector<int> &shape, void *input_data, void *output_data, bool enable_fp16) {
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   size_t dtype_size = enable_fp16 ? sizeof(float16_t) : sizeof(float);
   ocl_runtime->SetFp16Enable(enable_fp16);
@@ -124,7 +124,6 @@ void RunTestCaseMaxPooling(const std::vector<int> &shape, void *input_data, void
   }
 
   MS_LOG(INFO) << "Test MaxPool2d passed";
-  lite::opencl::OpenCLRuntime::DeleteInstance();
 }
 
 TEST_F(TestMaxPoolingOpenCL, MaxPoolingFp32) {

mindspore/lite/test/ut/src/runtime/kernel/opencl/prelu_tests.cc

@@ -77,7 +77,7 @@ TEST_F(TestPReluOpenCL, PReluFp32_dim4) {
   std::string weight_file = "/data/local/tmp/weight_data.bin";
   std::string standard_answer_file = "/data/local/tmp/caffe_prelu.bin";
   MS_LOG(INFO) << "-------------------->> Begin test PRelu!";
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   auto allocator = ocl_runtime->GetAllocator();
 
@@ -194,6 +194,5 @@ TEST_F(TestPReluOpenCL, PReluFp32_dim4) {
   delete weight_tensor;
   delete param;
   delete sub_graph;
-  lite::opencl::OpenCLRuntime::DeleteInstance();
 }
 }  // namespace mindspore

mindspore/lite/test/ut/src/runtime/kernel/opencl/reduce_tests.cc

@@ -31,7 +31,7 @@ class TestReduceOpenCL : public mindspore::CommonTest {
 
 void RunTestCaseReduce(const std::vector<int> &shape, void *input_data, void *output_data, bool enable_fp16,
                        int reduce_mode) {
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   size_t dtype_size = enable_fp16 ? sizeof(float16_t) : sizeof(float);
   ocl_runtime->SetFp16Enable(enable_fp16);
@@ -103,7 +103,6 @@ void RunTestCaseReduce(const std::vector<int> &shape, void *input_data, void *ou
   }
 
   MS_LOG(INFO) << "Test Reduce passed";
-  lite::opencl::OpenCLRuntime::DeleteInstance();
 }
 
 TEST_F(TestReduceOpenCL, ReduceMeanFp32) {

mindspore/lite/test/ut/src/runtime/kernel/opencl/reshape_tests.cc

@@ -31,7 +31,7 @@ class TestReshapeOpenCL : public mindspore::CommonTest {
 
 void RunTestCaseReshape(const std::vector<int> &shape, void *input_data, void *output_data, bool enable_fp16,
                         bool is_output_2d) {
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   size_t dtype_size = enable_fp16 ? sizeof(float16_t) : sizeof(float);
   ocl_runtime->SetFp16Enable(enable_fp16);
@@ -99,7 +99,6 @@ void RunTestCaseReshape(const std::vector<int> &shape, void *input_data, void *o
   }
 
   MS_LOG(INFO) << "Test Reshape passed";
-  lite::opencl::OpenCLRuntime::DeleteInstance();
 }
 
 TEST_F(TestReshapeOpenCL, ReshapeFp32) {

mindspore/lite/test/ut/src/runtime/kernel/opencl/scale_tests.cc

@@ -68,7 +68,7 @@ static void LogData(void *data, const int size, const std::string prefix) {
 template <class T>
 static void TestCase(const std::vector<int> &shape_a, const std::vector<int> &shape_b) {
   bool is_log_data = false;
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   auto allocator = ocl_runtime->GetAllocator();
 
   bool is_broadcast = shape_b.empty();
@@ -232,7 +232,6 @@ static void TestCase(const std::vector<int> &shape_a, const std::vector<int> &sh
   for (auto tensor : outputs) {
     delete tensor;
   }
-  lite::opencl::OpenCLRuntime::DeleteInstance();
 }
 
 class TestScaleOpenCL : public mindspore::CommonTest {

mindspore/lite/test/ut/src/runtime/kernel/opencl/slice_tests.cc

@@ -42,7 +42,7 @@ void CompareOutputData1(T *output_data, T *correct_data, int size, float err_bou
 
 TEST_F(TestSliceOpenCLfp32, Slicefp32CI) {
   MS_LOG(INFO) << " begin test ";
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   auto allocator = ocl_runtime->GetAllocator();
 
@@ -139,7 +139,6 @@ TEST_F(TestSliceOpenCLfp32, Slicefp32CI) {
 
   auto *output_data_gpu = reinterpret_cast<float *>(output_tensor->data_c());
   CompareOutputData1(output_data_gpu, correct_data, output_tensor->ElementsNum(), 0.0001);
-  lite::opencl::OpenCLRuntime::DeleteInstance();
   for (auto tensor : inputs) {
     tensor->SetData(nullptr);
     delete tensor;
@@ -153,7 +152,7 @@ TEST_F(TestSliceOpenCLfp32, Slicefp32CI) {
 
 TEST_F(TestSliceOpenCLfp32, Slicefp32input_dim4) {
   MS_LOG(INFO) << " begin test ";
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   auto allocator = ocl_runtime->GetAllocator();
 
@@ -248,7 +247,6 @@ TEST_F(TestSliceOpenCLfp32, Slicefp32input_dim4) {
 
   auto *output_data_gpu = reinterpret_cast<float *>(output_tensor->data_c());
   CompareOutputData1(output_data_gpu, correct_data, output_tensor->ElementsNum(), 0.0001);
-  lite::opencl::OpenCLRuntime::DeleteInstance();
   for (auto tensor : inputs) {
     tensor->SetData(nullptr);
     delete tensor;
@@ -262,7 +260,7 @@ TEST_F(TestSliceOpenCLfp32, Slicefp32input_dim4) {
 
 TEST_F(TestSliceOpenCLfp16, Slicefp16input_dim4) {
   MS_LOG(INFO) << " begin test ";
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->SetFp16Enable(true);
   ocl_runtime->Init();
   auto allocator = ocl_runtime->GetAllocator();
@@ -358,7 +356,6 @@ TEST_F(TestSliceOpenCLfp16, Slicefp16input_dim4) {
   sub_graph->Run();
   auto *output_data_gpu = reinterpret_cast<float16_t *>(output_tensor->data_c());
   CompareOutputData1(output_data_gpu, correct_data, output_tensor->ElementsNum(), 0.0001);
-  lite::opencl::OpenCLRuntime::DeleteInstance();
   for (auto tensor : inputs) {
     tensor->SetData(nullptr);
     delete tensor;

mindspore/lite/test/ut/src/runtime/kernel/opencl/softmax_tests.cc

@@ -30,7 +30,7 @@ class TestSoftmaxOpenCL : public mindspore::CommonTest {
 };
 
 void RunTestCaseSoftmax(const std::vector<int> &shape, void *input_data, void *output_data, bool enable_fp16) {
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   size_t dtype_size = enable_fp16 ? sizeof(float16_t) : sizeof(float);
   ocl_runtime->SetFp16Enable(enable_fp16);
@@ -103,7 +103,6 @@ void RunTestCaseSoftmax(const std::vector<int> &shape, void *input_data, void *o
   }
 
   MS_LOG(INFO) << "Test Softmax passed";
-  lite::opencl::OpenCLRuntime::DeleteInstance();
 }
 
 TEST_F(TestSoftmaxOpenCL, Softmax2DFp32) {

mindspore/lite/test/ut/src/runtime/kernel/opencl/to_format_tests.cc

@@ -29,7 +29,7 @@ class TestToFormatOpenCL : public mindspore::CommonTest {
 };
 
 TEST_F(TestToFormatOpenCL, ToFormatNHWC2NCHW) {
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   auto allocator = ocl_runtime->GetAllocator();
   int h = 64;
@@ -102,6 +102,5 @@ TEST_F(TestToFormatOpenCL, ToFormatNHWC2NCHW) {
   // compare
   CompareOutputData(output_data, correct_data, h * w * c, 0.00001);
   MS_LOG(INFO) << "Test TransposeFp32 passed";
-  lite::opencl::OpenCLRuntime::DeleteInstance();
 }
 }  // namespace mindspore

mindspore/lite/test/ut/src/runtime/kernel/opencl/transpose_tests.cc

@@ -30,7 +30,7 @@ class TestTransposeOpenCL : public mindspore::CommonTest {
 };
 
 void RunTestTranspose(const std::vector<int> &shape, void *input_data, void *output_data, bool enable_fp16) {
-  auto ocl_runtime = lite::opencl::OpenCLRuntime::GetInstance();
+  auto ocl_runtime = lite::opencl::OpenCLRuntimeWrapper().GetInstance();
   ocl_runtime->Init();
   size_t dtype_size = enable_fp16 ? sizeof(float16_t) : sizeof(float);
   ocl_runtime->SetFp16Enable(enable_fp16);
@@ -103,7 +103,6 @@ void RunTestTranspose(const std::vector<int> &shape, void *input_data, void *out
   }
 
   MS_LOG(INFO) << "Test TransposeFp32 passed";
-  lite::opencl::OpenCLRuntime::DeleteInstance();
 }
 
 TEST_F(TestTransposeOpenCL, TransposeNHWC2NCHWFp32) {