fix gpu matmul fp32 accuracy

mindspore/ccsrc/backend/kernel_compiler/gpu/math/matmul_gpu_kernel.h

@@ -42,7 +42,7 @@ class MatMulGpuKernel : public GpuKernel {
         dtype_a_(CUDA_R_32F),
         dtype_b_(CUDA_R_32F),
         dtype_c_(CUDA_R_32F),
-        algo_(CUBLAS_GEMM_DEFAULT_TENSOR_OP) {}
+        algo_(CUBLAS_GEMM_DEFAULT) {}
   ~MatMulGpuKernel() = default;
   const std::vector<size_t> &GetInputSizeList() const override { return input_size_list_; }
   const std::vector<size_t> &GetOutputSizeList() const override { return output_size_list_; }
@@ -85,6 +85,10 @@ class MatMulGpuKernel : public GpuKernel {
     dtype_a_ = GetCudaDataType(TypeIdLabel(AnfAlgo::GetInputDeviceDataType(kernel_node, 0)));
     dtype_b_ = GetCudaDataType(TypeIdLabel(AnfAlgo::GetInputDeviceDataType(kernel_node, 1)));
     dtype_c_ = GetCudaDataType(TypeIdLabel(AnfAlgo::GetOutputDeviceDataType(kernel_node, 0)));
+    if (dtype_a_ == CUDA_R_16F && dtype_b_ == CUDA_R_16F && dtype_c_ == CUDA_R_16F) {
+      MS_LOG(WARNING) << "input and output type is float16, allow to use Tensor Core operations if possible";
+      algo_ = CUBLAS_GEMM_DEFAULT_TENSOR_OP;
+    }
     auto output_shape = AnfAlgo::GetOutputInferShape(kernel_node, 0);
     is_null_input_ = CHECK_NULL_INPUT(output_shape);
     if (is_null_input_) {