cudnn 9.0 support... #25192
Labels
Comments
4 tasks
|
duplicate of #24983 There is no reason to target that support to 5.x. At least because it is applicable for 4.x too (no blockers are determined). Primary list of tasks for 5.0 is here: https://github.com/orgs/opencv/projects/1 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Describe the feature and motivation
[ 33%] Building CXX object modules/dnn/CMakeFiles/opencv_dnn.dir/src/layers/recurrent_layers.cpp.o cd ....../opencv/build/modules/dnn && /usr/bin/c++ -DCVAPI_EXPORTS -DCV_CUDA4DNN=1 -DENABLE_PLUGINS -DOPENCV_DNN_EXTERNAL_PROTOBUF=1 -DPROTOBUF_USE_DLLS -DVK_NO_PROTOTYPES -D_USE_MATH_DEFINES -D__OPENCV_BUILD=1 -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS -I....../opencv/build/3rdparty/ippicv/ippicv_lnx/icv/include -I....../opencv/build -I/usr/include/vulkan -I....../opencv/modules/dnn/include -I....../opencv/build/modules/dnn -I....../opencv_contrib/modules/cudev/include -I....../opencv/modules/core/include -I....../opencv/modules/imgproc/include -isystem /usr/include/eigen3/Eigen -isystem /usr/local/cuda/include -isystem /usr/include/eigen3 -fsigned-char -ffast-math -fno-finite-math-only -W -Wall -Wreturn-type -Wnon-virtual-dtor -Waddress -Wsequence-point -Wformat -Wformat-security -Winit-self -Wpointer-arith -Wsign-promo -Wuninitialized -Wsuggest-override -Wno-delete-non-virtual-dtor -Wno-comment -Wimplicit-fallthrough=3 -Wno-strict-overflow -fdiagnostics-show-option -Wno-long-long -pthread -fno-omit-frame-pointer -pg -g -msse -msse2 -msse3 -fvisibility=hidden -fvisibility-inlines-hidden -fopenmp -Wno-deprecated -Wno-missing-declarations -Wno-shadow -Wno-unused-parameter -Wno-sign-compare -Wno-undef -Wno-invalid-offsetof -O3 -DNDEBUG -DNDEBUG -std=c++17 -fPIC -MD -MT modules/dnn/CMakeFiles/opencv_dnn.dir/src/layers/recurrent_layers.cpp.o -MF CMakeFiles/opencv_dnn.dir/src/layers/recurrent_layers.cpp.o.d -o CMakeFiles/opencv_dnn.dir/src/layers/recurrent_layers.cpp.o -c ....../opencv/modules/dnn/src/layers/recurrent_layers.cpp In file included from ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../../cuda4dnn/csl/cudnn.hpp:8, from ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../../op_cuda.hpp:12, from ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/recurrent_cells.hpp:8, from ....../opencv/modules/dnn/src/layers/recurrent_layers.cpp:49: ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../csl/cudnn/recurrent.hpp: In instantiation of ‘cv::dnn::cuda4dnn::csl::cudnn::RNNDescriptor<T>::RNNDescriptor(const cv::dnn::cuda4dnn::csl::cudnn::Handle&, RNNMode, int, int, bool, const cv::dnn::cuda4dnn::csl::cudnn::DropoutDescriptor&) [with T = __half]’: ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../csl/tensor_ops.hpp:541:23: required from ‘cv::dnn::cuda4dnn::csl::LSTM<T>::LSTM(cv::dnn::cuda4dnn::csl::cudnn::Handle, const params_type&) [with T = __half]’ ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/recurrent_cells.hpp:48:21: required from ‘cv::dnn::cuda4dnn::LSTMOp<T>::LSTMOp(cv::dnn::cuda4dnn::csl::Stream, cv::dnn::cuda4dnn::csl::cudnn::Handle, const cv::Mat&, const cv::Mat&, const cv::Mat&, const cv::dnn::cuda4dnn::RNNConfiguration&) [with T = __half]’ ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../../op_cuda.hpp:194:37: required from ‘cv::Ptr<cv::dnn::dnn4_v20231225::BackendNode> cv::dnn::make_cuda_node(int, Args&& ...) [with NodeType = cuda4dnn::LSTMOp; Args = {cuda4dnn::csl::Stream, cuda4dnn::csl::cudnn::Handle, cv::Mat&, cv::Mat&, cv::Mat&, cuda4dnn::RNNConfiguration&}]’ ....../opencv/modules/dnn/src/layers/recurrent_layers.cpp:763:48: required from here ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../csl/cudnn/recurrent.hpp:122:58: error: ‘cudnnSetRNNDescriptor_v6’ was not declared in this scope; did you mean ‘cudnnSetRNNDescriptor_v8’? 122 | CUDA4DNN_CHECK_CUDNN(cudnnSetRNNDescriptor_v6( ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../../cuda4dnn/csl/cudnn/cudnn.hpp:22:53: note: in definition of macro ‘CUDA4DNN_CHECK_CUDNN’ 22 | ::cv::dnn::cuda4dnn::csl::cudnn::detail::check((call), CV_Func, __FILE__, __LINE__) | ^~~~ ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../csl/cudnn/recurrent.hpp: In instantiation of ‘size_t cv::dnn::cuda4dnn::csl::cudnn::getRNNWorkspaceSize(const Handle&, const RNNDescriptor<T>&, int, const TensorDescriptorsArray<T>&) [with T = __half; size_t = long unsigned int]’: ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../csl/tensor_ops.hpp:553:58: required from ‘cv::dnn::cuda4dnn::csl::LSTM<T>::LSTM(cv::dnn::cuda4dnn::csl::cudnn::Handle, const params_type&) [with T = __half]’ ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/recurrent_cells.hpp:48:21: required from ‘cv::dnn::cuda4dnn::LSTMOp<T>::LSTMOp(cv::dnn::cuda4dnn::csl::Stream, cv::dnn::cuda4dnn::csl::cudnn::Handle, const cv::Mat&, const cv::Mat&, const cv::Mat&, const cv::dnn::cuda4dnn::RNNConfiguration&) [with T = __half]’ ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../../op_cuda.hpp:194:37: required from ‘cv::Ptr<cv::dnn::dnn4_v20231225::BackendNode> cv::dnn::make_cuda_node(int, Args&& ...) [with NodeType = cuda4dnn::LSTMOp; Args = {cuda4dnn::csl::Stream, cuda4dnn::csl::cudnn::Handle, cv::Mat&, cv::Mat&, cv::Mat&, cuda4dnn::RNNConfiguration&}]’ ....../opencv/modules/dnn/src/layers/recurrent_layers.cpp:763:48: required from here ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../csl/cudnn/recurrent.hpp:166:50: error: ‘cudnnGetRNNWorkspaceSize’ was not declared in this scope; did you mean ‘cudnnGetRNNWeightSpaceSize’? 166 | CUDA4DNN_CHECK_CUDNN(cudnnGetRNNWorkspaceSize(handle.get(), rnnDesc.get(), seqLength, ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../../cuda4dnn/csl/cudnn/cudnn.hpp:22:53: note: in definition of macro ‘CUDA4DNN_CHECK_CUDNN’ 22 | ::cv::dnn::cuda4dnn::csl::cudnn::detail::check((call), CV_Func, __FILE__, __LINE__) | ^~~~ ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../csl/cudnn/recurrent.hpp: In instantiation of ‘cv::dnn::cuda4dnn::csl::cudnn::RNNDescriptor<T>::RNNDescriptor(const cv::dnn::cuda4dnn::csl::cudnn::Handle&, RNNMode, int, int, bool, const cv::dnn::cuda4dnn::csl::cudnn::DropoutDescriptor&) [with T = float]’: ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../csl/tensor_ops.hpp:541:23: required from ‘cv::dnn::cuda4dnn::csl::LSTM<T>::LSTM(cv::dnn::cuda4dnn::csl::cudnn::Handle, const params_type&) [with T = float]’ ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/recurrent_cells.hpp:48:21: required from ‘cv::dnn::cuda4dnn::LSTMOp<T>::LSTMOp(cv::dnn::cuda4dnn::csl::Stream, cv::dnn::cuda4dnn::csl::cudnn::Handle, const cv::Mat&, const cv::Mat&, const cv::Mat&, const cv::dnn::cuda4dnn::RNNConfiguration&) [with T = float]’ ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../../op_cuda.hpp:196:37: required from ‘cv::Ptr<cv::dnn::dnn4_v20231225::BackendNode> cv::dnn::make_cuda_node(int, Args&& ...) [with NodeType = cuda4dnn::LSTMOp; Args = {cuda4dnn::csl::Stream, cuda4dnn::csl::cudnn::Handle, cv::Mat&, cv::Mat&, cv::Mat&, cuda4dnn::RNNConfiguration&}]’ ....../opencv/modules/dnn/src/layers/recurrent_layers.cpp:763:48: required from here ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../csl/cudnn/recurrent.hpp:122:58: error: ‘cudnnSetRNNDescriptor_v6’ was not declared in this scope; did you mean ‘cudnnSetRNNDescriptor_v8’? 122 | CUDA4DNN_CHECK_CUDNN(cudnnSetRNNDescriptor_v6( ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../../cuda4dnn/csl/cudnn/cudnn.hpp:22:53: note: in definition of macro ‘CUDA4DNN_CHECK_CUDNN’ 22 | ::cv::dnn::cuda4dnn::csl::cudnn::detail::check((call), CV_Func, __FILE__, __LINE__) | ^~~~ ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../csl/cudnn/recurrent.hpp: In instantiation of ‘size_t cv::dnn::cuda4dnn::csl::cudnn::getRNNWorkspaceSize(const Handle&, const RNNDescriptor<T>&, int, const TensorDescriptorsArray<T>&) [with T = float; size_t = long unsigned int]’: ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../csl/tensor_ops.hpp:553:58: required from ‘cv::dnn::cuda4dnn::csl::LSTM<T>::LSTM(cv::dnn::cuda4dnn::csl::cudnn::Handle, const params_type&) [with T = float]’ ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/recurrent_cells.hpp:48:21: required from ‘cv::dnn::cuda4dnn::LSTMOp<T>::LSTMOp(cv::dnn::cuda4dnn::csl::Stream, cv::dnn::cuda4dnn::csl::cudnn::Handle, const cv::Mat&, const cv::Mat&, const cv::Mat&, const cv::dnn::cuda4dnn::RNNConfiguration&) [with T = float]’ ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../../op_cuda.hpp:196:37: required from ‘cv::Ptr<cv::dnn::dnn4_v20231225::BackendNode> cv::dnn::make_cuda_node(int, Args&& ...) [with NodeType = cuda4dnn::LSTMOp; Args = {cuda4dnn::csl::Stream, cuda4dnn::csl::cudnn::Handle, cv::Mat&, cv::Mat&, cv::Mat&, cuda4dnn::RNNConfiguration&}]’ ....../opencv/modules/dnn/src/layers/recurrent_layers.cpp:763:48: required from here ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../csl/cudnn/recurrent.hpp:166:50: error: ‘cudnnGetRNNWorkspaceSize’ was not declared in this scope; did you mean ‘cudnnGetRNNWeightSpaceSize’? 166 | CUDA4DNN_CHECK_CUDNN(cudnnGetRNNWorkspaceSize(handle.get(), rnnDesc.get(), seqLength, ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../../cuda4dnn/csl/cudnn/cudnn.hpp:22:53: note: in definition of macro ‘CUDA4DNN_CHECK_CUDNN’ 22 | ::cv::dnn::cuda4dnn::csl::cudnn::detail::check((call), CV_Func, __FILE__, __LINE__) | ^~~~ ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../csl/cudnn/recurrent.hpp: In instantiation of ‘void cv::dnn::cuda4dnn::csl::cudnn::LSTMForward(const Handle&, const RNNDescriptor<T>&, const FilterDescriptor<T>&, cv::dnn::cuda4dnn::csl::DevicePtr<const T>, const TensorDescriptorsArray<T>&, cv::dnn::cuda4dnn::csl::DevicePtr<const T>, const TensorDescriptor<T>&, cv::dnn::cuda4dnn::csl::DevicePtr<const T>, const TensorDescriptor<T>&, cv::dnn::cuda4dnn::csl::DevicePtr<const T>, int, const TensorDescriptorsArray<T>&, cv::dnn::cuda4dnn::csl::DevicePtr<T>, cv::dnn::cuda4dnn::csl::DevicePtr<T>, cv::dnn::cuda4dnn::csl::WorkspaceInstance) [with T = float]’: ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../csl/tensor_ops.hpp:563:34: required from ‘void cv::dnn::cuda4dnn::csl::LSTM<T>::inference(cv::dnn::cuda4dnn::csl::TensorView<T>, cv::dnn::cuda4dnn::csl::TensorSpan<T>, cv::dnn::cuda4dnn::csl::TensorSpan<T>, cv::dnn::cuda4dnn::csl::TensorView<T>, cv::dnn::cuda4dnn::csl::TensorView<T>, cv::dnn::cuda4dnn::csl::TensorView<T>, cv::dnn::cuda4dnn::csl::WorkspaceInstance) [with T = float]’ ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/recurrent_cells.hpp:79:23: required from ‘void cv::dnn::cuda4dnn::LSTMOp<T>::forward(const std::vector<cv::Ptr<cv::dnn::dnn4_v20231225::BackendWrapper> >&, const std::vector<cv::Ptr<cv::dnn::dnn4_v20231225::BackendWrapper> >&, cv::dnn::cuda4dnn::csl::Workspace&) [with T = float]’ ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/recurrent_cells.hpp:63:10: required from here ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../csl/cudnn/recurrent.hpp:182:50: error: ‘cudnnRNNForwardInference’ was not declared in this scope 182 | CUDA4DNN_CHECK_CUDNN(cudnnRNNForwardInference(handle.get(), rnnDesc.get(), seqLength, ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../../cuda4dnn/csl/cudnn/cudnn.hpp:22:53: note: in definition of macro ‘CUDA4DNN_CHECK_CUDNN’ 22 | ::cv::dnn::cuda4dnn::csl::cudnn::detail::check((call), CV_Func, __FILE__, __LINE__) | ^~~~ ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../csl/cudnn/recurrent.hpp: In instantiation of ‘void cv::dnn::cuda4dnn::csl::cudnn::LSTMForward(const Handle&, const RNNDescriptor<T>&, const FilterDescriptor<T>&, cv::dnn::cuda4dnn::csl::DevicePtr<const T>, const TensorDescriptorsArray<T>&, cv::dnn::cuda4dnn::csl::DevicePtr<const T>, const TensorDescriptor<T>&, cv::dnn::cuda4dnn::csl::DevicePtr<const T>, const TensorDescriptor<T>&, cv::dnn::cuda4dnn::csl::DevicePtr<const T>, int, const TensorDescriptorsArray<T>&, cv::dnn::cuda4dnn::csl::DevicePtr<T>, cv::dnn::cuda4dnn::csl::DevicePtr<T>, cv::dnn::cuda4dnn::csl::WorkspaceInstance) [with T = __half]’: ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../csl/tensor_ops.hpp:563:34: required from ‘void cv::dnn::cuda4dnn::csl::LSTM<T>::inference(cv::dnn::cuda4dnn::csl::TensorView<T>, cv::dnn::cuda4dnn::csl::TensorSpan<T>, cv::dnn::cuda4dnn::csl::TensorSpan<T>, cv::dnn::cuda4dnn::csl::TensorView<T>, cv::dnn::cuda4dnn::csl::TensorView<T>, cv::dnn::cuda4dnn::csl::TensorView<T>, cv::dnn::cuda4dnn::csl::WorkspaceInstance) [with T = __half]’ ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/recurrent_cells.hpp:79:23: required from ‘void cv::dnn::cuda4dnn::LSTMOp<T>::forward(const std::vector<cv::Ptr<cv::dnn::dnn4_v20231225::BackendWrapper> >&, const std::vector<cv::Ptr<cv::dnn::dnn4_v20231225::BackendWrapper> >&, cv::dnn::cuda4dnn::csl::Workspace&) [with T = __half]’ ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/recurrent_cells.hpp:63:10: required from here ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../csl/cudnn/recurrent.hpp:182:50: error: ‘cudnnRNNForwardInference’ was not declared in this scope 182 | CUDA4DNN_CHECK_CUDNN(cudnnRNNForwardInference(handle.get(), rnnDesc.get(), seqLength, ....../opencv/modules/dnn/src/layers/../cuda4dnn/primitives/../../cuda4dnn/csl/cudnn/cudnn.hpp:22:53: note: in definition of macro ‘CUDA4DNN_CHECK_CUDNN’ 22 | ::cv::dnn::cuda4dnn::csl::cudnn::detail::check((call), CV_Func, __FILE__, __LINE__) | ^~~~ make[2]: *** [modules/dnn/CMakeFiles/opencv_dnn.dir/build.make:10634: modules/dnn/CMakeFiles/opencv_dnn.dir/src/layers/recurrent_layers.cpp.o] Error 1 make[2]: Leaving directory '....../opencv/build' make[1]: *** [CMakeFiles/Makefile2:4141: modules/dnn/CMakeFiles/opencv_dnn.dir/all] Error 2 make[1]: Leaving directory '....../opencv/build' make: *** [Makefile:169: all] Error 2Additional context
When is your timeline for opencv-5.0?
opencv-5.0 will be compatible with cudnn 9.0, right?
The text was updated successfully, but these errors were encountered: