I have successfully installed caffe with CUDA 6.5. But when I install it with cuDNN 6.5 v2 after uncommenting USE_CUDNN:=1, make all command gives the followin error.
In file included from ./include/caffe/loss_layers.hpp:11:0,
from ./include/caffe/common_layers.hpp:12,
from ./include/caffe/vision_layers.hpp:10,
from src/caffe/layer_factory.cpp:6:
./include/caffe/neuron_layers.hpp:434:3: error: ‘cudnnHandle_t’ does not name a type
cudnnHandle_t handle_;
^
./include/caffe/neuron_layers.hpp:435:3: error: ‘cudnnTensorDescriptor_t’ does not name a type
cudnnTensorDescriptor_t bottom_desc_;
^
./include/caffe/neuron_layers.hpp:436:3: error: ‘cudnnTensorDescriptor_t’ does not name a type
cudnnTensorDescriptor_t top_desc_;
^
./include/caffe/neuron_layers.hpp:517:3: error: ‘cudnnHandle_t’ does not name a type
cudnnHandle_t handle_;
^
./include/caffe/neuron_layers.hpp:518:3: error: ‘cudnnTensorDescriptor_t’ does not name a type
cudnnTensorDescriptor_t bottom_desc_;
^
./include/caffe/neuron_layers.hpp:519:3: error: ‘cudnnTensorDescriptor_t’ does not name a type
cudnnTensorDescriptor_t top_desc_;
^
./include/caffe/neuron_layers.hpp:602:3: error: ‘cudnnHandle_t’ does not name a type
cudnnHandle_t handle_;
^
./include/caffe/neuron_layers.hpp:603:3: error: ‘cudnnTensorDescriptor_t’ does not name a type
cudnnTensorDescriptor_t bottom_desc_;
^
./include/caffe/neuron_layers.hpp:604:3: error: ‘cudnnTensorDescriptor_t’ does not name a type
cudnnTensorDescriptor_t top_desc_;
^
In file included from ./include/caffe/vision_layers.hpp:10:0,
from src/caffe/layer_factory.cpp:6:
./include/caffe/common_layers.hpp:426:3: error: ‘cudnnHandle_t’ does not name a type
cudnnHandle_t handle_;
^
./include/caffe/common_layers.hpp:427:3: error: ‘cudnnTensorDescriptor_t’ does not name a type
cudnnTensorDescriptor_t bottom_desc_;
^
./include/caffe/common_layers.hpp:428:3: error: ‘cudnnTensorDescriptor_t’ does not name a type
cudnnTensorDescriptor_t top_desc_;
^
In file included from src/caffe/layer_factory.cpp:6:0:
./include/caffe/vision_layers.hpp:247:3: error: ‘cudnnHandle_t’ does not name a type
cudnnHandle_t* handle_;
^
./include/caffe/vision_layers.hpp:248:3: error: ‘cudaStream_t’ does not name a type
cudaStream_t* stream_;
^
./include/caffe/vision_layers.hpp:249:10: error: ‘cudnnTensorDescriptor_t’ was not declared in this scope
vector
^
./include/caffe/vision_layers.hpp:249:33: error: template argument 1 is invalid
vector
^
./include/caffe/vision_layers.hpp:249:33: error: template argument 2 is invalid
./include/caffe/vision_layers.hpp:250:3: error: ‘cudnnTensorDescriptor_t’ does not name a type
cudnnTensorDescriptor_t bias_desc_;
^
./include/caffe/vision_layers.hpp:251:3: error: ‘cudnnFilterDescriptor_t’ does not name a type
cudnnFilterDescriptor_t filter_desc_;
^
./include/caffe/vision_layers.hpp:252:10: error: ‘cudnnConvolutionDescriptor_t’ was not declared in this scope
vector
^
./include/caffe/vision_layers.hpp:252:38: error: template argument 1 is invalid
vector
^
./include/caffe/vision_layers.hpp:252:38: error: template argument 2 is invalid
./include/caffe/vision_layers.hpp:449:3: error: ‘cudnnHandle_t’ does not name a type
cudnnHandle_t handle_;
^
./include/caffe/vision_layers.hpp:450:3: error: ‘cudnnTensorDescriptor_t’ does not name a type
cudnnTensorDescriptor_t bottom_desc_, top_desc_;
^
./include/caffe/vision_layers.hpp:451:3: error: ‘cudnnPoolingDescriptor_t’ does not name a type
cudnnPoolingDescriptor_t pooling_desc_;
^
./include/caffe/vision_layers.hpp:452:3: error: ‘cudnnPoolingMode_t’ does not name a type
cudnnPoolingMode_t mode_;
^
src/caffe/layer_factory.cpp: In function ‘boost::shared_ptrcaffe::Layer
}
^
src/caffe/layer_factory.cpp: In function ‘boost::shared_ptrcaffe::Layer
}
^
src/caffe/layer_factory.cpp: In function ‘boost::shared_ptrcaffe::Layer
}
^
src/caffe/layer_factory.cpp: In function ‘boost::shared_ptrcaffe::Layer
}
^
src/caffe/layer_factory.cpp: In function ‘boost::shared_ptrcaffe::Layer
}
^
src/caffe/layer_factory.cpp: In function ‘boost::shared_ptrcaffe::Layer
}
^
src/caffe/layer_factory.cpp: In function ‘boost::shared_ptrcaffe::Layer
}
^
src/caffe/layer_factory.cpp: In function ‘boost::shared_ptrcaffe::Layer
}
^
src/caffe/layer_factory.cpp: In function ‘boost::shared_ptrcaffe::Layer
}
^
src/caffe/layer_factory.cpp: In function ‘boost::shared_ptrcaffe::Layer
}
^
src/caffe/layer_factory.cpp: In function ‘boost::shared_ptrcaffe::Layer
}
^
src/caffe/layer_factory.cpp: In function ‘boost::shared_ptrcaffe::Layer
}
^
make: *** [.build_release/src/caffe/layer_factory.o] Error 1
abhi9git
All 12 comments
This tracker is for issues and bugs with Caffe; please ask installation and other usage questions on the caffe-users list. Thanks!
longjon
on 17 May 2015
This is a bug with Caffe. cudnnTensorDescriptor_t doesn't exist in the header file of cuDNN v2. It has been renamed to cudnnTensor4dDescriptor_t.
futurely
on 20 May 2015
Sorry, this is a false alarm. I just noticed that the error was caused by cuDNN R1(v1). Newly downloaded v2 works very well.
futurely
on 20 May 2015
@futurely
Even if I used the cuDNN-v2(cudnn-6.5-linux-x64-v2.tgz from nvidia website) with the latest Caffe from this git repository, the problem is not resolved.
Jeongseob
on 18 Jun 2015
@Jeongseob The filenames of cuDNN v1 and v2 are the same. Please delete the existing cuDNN header file and shared library and copy the v2 files again.
futurely
on 19 Jun 2015
@futurely Actually, I installed the cuDNN-v2 library on my home directory and just exported the LD_LIBRARY_PATH. I did not install the cuDNN v1.
Jeongseob
on 19 Jun 2015
It's much simpler to install them along with CUDA by
cd <cuDNN path>
sudo cp cudnn.h /usr/local/cuda/include
sudo cp libcudnn* /usr/local/cuda/lib64
## /usr/local/cuda/lib64 should have been added to LD_LIBRARY_PATH
sudo ldconfig
@futurely Do I have to install the cuDNN on the "/usr/local/cuda/" path? Actually, I do not want to install the cuDNN on the system-wide. Is there any problem if I just set the path to my own home directory?
Jeongseob
on 19 Jun 2015
There's no difference between using them with using any other libraries.
futurely
on 22 Jun 2015
got this problem too, I used cuDNN v2 with latest caffe from the master. These are errors:
(If I use cuDNN v1, all will be compiled successfully)
CXX/LD -o .build_release/examples/mnist/convert_mnist_data.bin
.build_release/lib/libcaffe.so: undefined reference to `cudnnSetFilter4dDescriptor'
.build_release/lib/libcaffe.so: undefined reference to `cudnnGetConvolutionForwardAlgorithm'
.build_release/lib/libcaffe.so: undefined reference to `cv::imread(cv::String const&, int)'
.build_release/lib/libcaffe.so: undefined reference to `cudnnCreateTensorDescriptor'
.build_release/lib/libcaffe.so: undefined reference to `cv::imencode(cv::String const&, cv::_InputArray const&, std::vector<unsigned char, std::allocator<unsigned char> >&, std::vector<int, std::allocator<int> > const&)'
.build_release/lib/libcaffe.so: undefined reference to `cudnnDestroyTensorDescriptor'
.build_release/lib/libcaffe.so: undefined reference to `cudnnGetConvolutionForwardWorkspaceSize'
.build_release/lib/libcaffe.so: undefined reference to `cudnnAddTensor'
.build_release/lib/libcaffe.so: undefined reference to `cudnnSetPooling2dDescriptor'
.build_release/lib/libcaffe.so: undefined reference to `cv::imdecode(cv::_InputArray const&, int)'
.build_release/lib/libcaffe.so: undefined reference to `cudnnSetConvolution2dDescriptor'
collect2: error: ld returned 1 exit status
make: *** [.build_release/tools/convert_imageset.bin] Error 1
make: *** Waiting for unfinished jobs....
LiberiFatali
on 31 Jul 2015
I solved the problem above by replacing old cuDNN files with v2 files in every paths that contain libcudnn*. In my case (Ubuntu 14.04):
user_Ubuntu:~/caffe-master$ locate libcudnn_static.a
/home/user/cudnn-6.5-linux-x64-v2/libcudnn_static.a
/lib64/libcudnn_static.a
/usr/lib/libcudnn/libcudnn_static.a
/usr/local/cuda-7.0/targets/x86_64-linux/lib/libcudnn_static.a
/usr/local/lib/libcudnn_static.a
So I did:
cd <cuDNN path>
sudo cp cudnn.h /usr/local/cuda/include
sudo cp libcudnn* /usr/local/cuda/lib64
# Add /usr/local/cuda/lib64 to LD_LIBRARY_PATH in ~/.bashrc
sudo cp libcudnn* /lib64/
sudo cp libcudnn* /usr/lib/libcudnn/
sudo cp libcudnn* /usr/local/cuda-7.0/targets/x86_64-linux/lib/
sudo cp libcudnn* /usr/local/lib
sudo ldconfig
Then caffe-master is compiled succesfully.
LiberiFatali
on 31 Jul 2015
mark
ysh329
on 23 Jun 2018
Related issues
inferrna
·
3Comments
LarsHH
·
3Comments
amit-dat
·
3Comments
Ruhjkg
·
3Comments
greatgao
·
3Comments
Most helpful comment
I solved the problem above by replacing old cuDNN files with v2 files in every paths that contain libcudnn*. In my case (Ubuntu 14.04):
So I did:
Then caffe-master is compiled succesfully.