Nigel
/
mindspore
forked from mindspore-Ecosystem/mindspore
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

!6458 GPU codex fix 

Merge pull request !6458 from VectorSL/codex-fix
This commit is contained in:
mindspore-ci-bot 2020-09-18 15:20:22 +08:00 committed by Gitee
parent 07e942c2da 9e6bd72e04
commit 076d8ae530
3 changed files with 2 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
mindspore/ccsrc/backend/kernel_compiler/gpu/math/matmul_gpu_kernel.h
View File

@ -86,7 +86,7 @@ class MatMulGpuKernel : public GpuKernel {
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";
MS_LOG(INFO) << "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);

2
mindspore/ccsrc/backend/kernel_compiler/gpu/nn/pooling_grad_gpu_kernel.h
View File

@ -86,6 +86,7 @@ class PoolingGradGpuKernel : public GpuKernel {
auto dout_shape = AnfAlgo::GetInputDeviceShape(kernel_node, 2);
auto output_shape = AnfAlgo::GetOutputDeviceShape(kernel_node, 0);
data_format_ = AnfAlgo::GetInputFormat(kernel_node, 0);
cudnn_data_type_ = GetCudnnDataType(TypeIdLabel(AnfAlgo::GetInputDeviceDataType(kernel_node, 0)));
is_null_input_ = CHECK_NULL_INPUT(input_shape) || CHECK_NULL_INPUT(input_mask);
if (is_null_input_) {
MS_LOG(WARNING) << "PoolingGradGpuKernel input is null.";
@ -204,7 +205,6 @@ class PoolingGradGpuKernel : public GpuKernel {
"cudnnSetPoolingNdDescriptor failed");
}
void SetPoolingMode(const CNodePtr &kernel_node) {
cudnn_data_type_ = GetCudnnDataType(TypeIdLabel(AnfAlgo::GetInputDeviceDataType(kernel_node, 0)));
mode_ = AnfAlgo::GetCNodeName(kernel_node);
if (mode_ == "AvgPoolGradGpu") {
pooling_mode_ = CUDNN_POOLING_AVERAGE_COUNT_INCLUDE_PADDING;

4
mindspore/ccsrc/runtime/device/gpu/kernel_info_setter.cc
View File

@ -345,7 +345,6 @@ void SetKernelInfo(const CNodePtr &kernel_node, KernelType kernel_type) {
SetGraphKernelInfo(kernel_node, func_graph);
return;
}
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) {
@ -368,12 +367,10 @@ void SetKernelInfo(const CNodePtr &kernel_node, KernelType kernel_type) {
builder->SetInputsDeviceType(inputs_type);
builder->SetOutputsFormat(outputs_format);
builder->SetOutputsDeviceType(outputs_type);
bool result = false;
if (kernel_type == UNKNOWN_KERNEL_TYPE) {
result =
kernel::GpuKernelFactory::GetInstance().SearchRegistered(AnfAlgo::GetCNodeName(kernel_node), builder->Build());
if (!result) {
result = SelectAkgKernel(kernel_node, builder->Build());
kernel_type = AKG_KERNEL;
@ -381,7 +378,6 @@ void SetKernelInfo(const CNodePtr &kernel_node, KernelType kernel_type) {
} else if (kernel_type == AKG_KERNEL) {
result = SelectAkgKernel(kernel_node, builder->Build());
}
if (!result) {
PrintUnsupportedTypeException(kernel_node, inputs_type, outputs_type);
return;