!4205 transpose output set to image2d and add mobilenetv2 GPU testcase

Merge pull request !4205 from chenzupeng/master-lite

mindspore/lite/src/runtime/kernel/opencl/cl/fp16/transpose.cl

@@ -1,8 +1,9 @@
 #define FLT half
 #define FLT4 half4
 #define READ_IMAGE read_imageh
+#define WRITE_IMAGE write_imageh
 __constant sampler_t smp_zero = CLK_NORMALIZED_COORDS_FALSE | CLK_ADDRESS_CLAMP | CLK_FILTER_NEAREST;
-__kernel void transpose(__read_only image2d_t src_data, __global float4 *dst_data, int2 HW, int2 C) {
+__kernel void transpose(__read_only image2d_t src_data, __write_only image2d_t dst_data, int2 HW, int2 C) {
   int X = get_global_id(0);
   int Y = get_global_id(1);
   if (X >= HW.y || Y >= C.y) {
@@ -37,8 +38,8 @@ __kernel void transpose(__read_only image2d_t src_data, __global float4 *dst_dat
   result[3].z = x2.w;
   result[3].w = x3.w;
 
-  dst_data[4 * Y * HW.y + X] = result[0];
+  WRITE_IMAGE(dst_data, (int2)(X, 4 * Y), result[0]);
-  dst_data[(4 * Y + 1) * HW.y + X] = result[1];
+  WRITE_IMAGE(dst_data, (int2)(X, 4 * Y + 1), result[1]);
-  dst_data[(4 * Y + 2) * HW.y + X] = result[2];
+  WRITE_IMAGE(dst_data, (int2)(X, 4 * Y + 2), result[2]);
-  dst_data[(4 * Y + 3) * HW.y + X] = result[3];
+  WRITE_IMAGE(dst_data, (int2)(X, 4 * Y + 3), result[3]);
 }

mindspore/lite/src/runtime/kernel/opencl/cl/fp32/transpose.cl

@@ -1,8 +1,9 @@
 #define FLT float
 #define FLT4 float4
 #define READ_IMAGE read_imagef
+#define WRITE_IMAGE write_imagef
 __constant sampler_t smp_zero = CLK_NORMALIZED_COORDS_FALSE | CLK_ADDRESS_CLAMP | CLK_FILTER_NEAREST;
-__kernel void transpose(__read_only image2d_t src_data, __global float4 *dst_data, int2 HW, int2 C) {
+__kernel void transpose(__read_only image2d_t src_data, __write_only image2d_t dst_data, int2 HW, int2 C) {
   int X = get_global_id(0);
   int Y = get_global_id(1);
   if (X >= HW.y || Y >= C.y) {
@@ -37,8 +38,8 @@ __kernel void transpose(__read_only image2d_t src_data, __global float4 *dst_dat
   result[3].z = x2.w;
   result[3].w = x3.w;
 
-  dst_data[4 * Y * HW.y + X] = result[0];
+  WRITE_IMAGE(dst_data, (int2)(X, 4 * Y), result[0]);
-  dst_data[(4 * Y + 1) * HW.y + X] = result[1];
+  WRITE_IMAGE(dst_data, (int2)(X, 4 * Y + 1), result[1]);
-  dst_data[(4 * Y + 2) * HW.y + X] = result[2];
+  WRITE_IMAGE(dst_data, (int2)(X, 4 * Y + 2), result[2]);
-  dst_data[(4 * Y + 3) * HW.y + X] = result[3];
+  WRITE_IMAGE(dst_data, (int2)(X, 4 * Y + 3), result[3]);
 }

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

@@ -43,7 +43,7 @@ void ArithmeticOpenCLKernel::Image2dGetWorkGroupSize() {
   size_t H = outputs_[0]->Batch() * outputs_[0]->Height();
   size_t W = outputs_[0]->Width() * UP_DIV(outputs_[0]->Channel(), C4NUM);
   local_size_ = {16, 16};
-  global_size_ = {H, W};
+  global_size_ = {W, H};
 }
 
 void ArithmeticOpenCLKernel::BufferGetWorkGroupSize() {
@@ -140,7 +140,7 @@ int ArithmeticOpenCLKernel::Run() {
         bias_ = -1 * value;
         break;
       case PrimitiveType_Div:
-        bias_ = 1 / value;
+        weight_ = 1 / value;
         break;
       default:
         MS_LOG(ERROR) << "Error Operator type " << opParameter->type_;
@@ -152,7 +152,7 @@ int ArithmeticOpenCLKernel::Run() {
   runtime_->SetKernelArg(kernel_, arg_idx++, outputs_[0]->Data());
   int H = outputs_[0]->Batch() * outputs_[0]->Height();
   int W = outputs_[0]->Width() * UP_DIV(outputs_[0]->Channel(), C4NUM);
-  cl_int2 output_shape{H, W};
+  cl_int2 output_shape{W, H};
   runtime_->SetKernelArg(kernel_, arg_idx++, output_shape);
   runtime_->RunKernel(kernel_, global_size_, local_size_, nullptr);
   return 0;

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

@@ -67,6 +67,21 @@ int TransposeOpenCLKernel::Init() {
 
 int TransposeOpenCLKernel::ReSize() { return 0; }
 
+int TransposeOpenCLKernel::GetImageSize(size_t idx, std::vector<size_t> *img_size) {
+  size_t im_dst_x, im_dst_y;
+  im_dst_x = UP_DIV(outputs_[0]->Height() * outputs_[0]->Width(), C4NUM);
+  im_dst_y = outputs_[0]->Channel();
+#ifdef ENABLE_FP16
+  size_t img_dtype = CL_HALF_FLOAT;
+#else
+  size_t img_dtype = CL_FLOAT;
+#endif
+  img_size->clear();
+  std::vector<size_t> vec{im_dst_x, im_dst_y, img_dtype};
+  *img_size = vec;
+  return RET_OK;
+}
+
 int TransposeOpenCLKernel::Run() {
   MS_LOG(DEBUG) << this->Name() << " Running!";
   std::vector<int> shapex = inputs_[0]->shape();

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

@@ -35,7 +35,7 @@ class TransposeOpenCLKernel : public OpenCLKernel {
   int Init() override;
   int ReSize() override;
   int Run() override;
-
+  int GetImageSize(size_t idx, std::vector<size_t> *img_size) override;
  private:
   cl::Kernel kernel_;
 };

mindspore/lite/test/st/benchmark_test.cc

@@ -50,6 +50,15 @@ auto status = RunBenchmark(5, argv);
 ASSERT_EQ(status, RET_OK);
 }
 
+TEST_F(BenchmarkTest, Test_MV2_GPU) {
+const char *argv[] = {"./benchmark", "--modelPath=./hiai/mobilenet_v2.ms",
+                                     "--inDataPath=./hiai/mobilenet_v2_in.bin",
+                                     "--calibDataPath=./hiai/mobilenet_v2_out.bin",
+                                     "--device=GPU"};
+auto status = RunBenchmark(5, argv);
+ASSERT_EQ(status, RET_OK);
+}
+
 TEST_F(BenchmarkTest, TestHebing) {
   const char *argv[] = {"./benchmark", "--modelPath=./hiai/model_hebing_3branch.ms",
                                        "--inDataPath=./hiai/model_hebing_3branch.bin",

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

@@ -66,6 +66,9 @@ TEST_F(TestTransposeOpenCL, TransposeFp32) {
   size_n = size_n > 100 ? 100 : size_n;
   for (int i = 0; i < size_n; i++) {
     std::cout << output_data[i] << " ";
+    if ((i + 1) % c == 0) {
+      std::cout << std::endl;
+    }
   }
   std::cout << std::endl;