-
What is CUDA? CUDA is a parallel computing platform and programming model that makes using a GPU for general purpose computing simple. The CUDA platform is a software layer that gives direct access to the GPU's virtual instruction set and parallel computational elements, for the execution of compute kernels
-
What is CUDNN? The NVIDIA CUDA® Deep Neural Network library (cuDNN) is a GPU-accelerated library of primitives for deep neural networks. cuDNN provides highly tuned implementations for standard routines such as forward and backward convolution, pooling, normalization, and activation layers. Built on CUDA
-
What is CUDA Toolkit? (installs Driver and CUDA) The toolkit includes GPU-accelerated libraries, debugging and optimization tools, a C/C++ compiler and a runtime library
driver version - cat /proc/driver/nvidia/version nvidia-smi
cuda version - cat /usr/local/cuda/version.txt nvidia-smi
cudnn version nvcc --version
Download links: https://developer.nvidia.com/cuda-10.0-download-archive?target_os=Linux https://developer.nvidia.com/rdp/cudnn-download
Official documentation: https://docs.nvidia.com/deeplearning/sdk/cudnn-install/index.html#installdriver
sudo service lightdm stop # or pkill Xorg
install cuda toolkit install cudnn
sudo modprobe -r nvidia sudo modprobe nvidia
sudo service lightdm start
! Only the driver must be installed on the host. The CUDA version from the container must be compatible with the host driver.
sudo apt-get install nvidia-container-toolkit
sudo systemctl daemon-reload sudo systemctl restart docker