CUDA: Difference between revisions
(→Linux) |
(→Linux) |
||
Line 4: | Line 4: | ||
[https://www.pugetsystems.com/labs/hpc/How-To-Install-CUDA-10-1-on-Ubuntu-19-04-1405/#Step3)InstallCUDA\ Reference] | [https://www.pugetsystems.com/labs/hpc/How-To-Install-CUDA-10-1-on-Ubuntu-19-04-1405/#Step3)InstallCUDA\ Reference] | ||
* | * Install the latest nvidia drivers from the standard repo, e.g. <code>nvidia-drivers-440</code><br> | ||
* Install [https://developer.nvidia.com/cuda-toolkit Cuda Toolkit] separately without the drivers.<br> | * Install [https://developer.nvidia.com/cuda-toolkit Cuda Toolkit] separately without the drivers.<br> | ||
** Use one of the deb install options. | ** Use one of the deb install options. | ||
Line 10: | Line 10: | ||
** [https://developer.nvidia.com/rdp/cudnn-download cuDnn]<br> | ** [https://developer.nvidia.com/rdp/cudnn-download cuDnn]<br> | ||
** TensorRT | ** TensorRT | ||
;Copied from tensorflow | |||
<pre> | |||
# Add NVIDIA package repositories | |||
wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/cuda-repo-ubuntu1804_10.1.243-1_amd64.deb | |||
sudo apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/7fa2af80.pub | |||
sudo dpkg -i cuda-repo-ubuntu1804_10.1.243-1_amd64.deb | |||
sudo apt-get update | |||
wget http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1804/x86_64/nvidia-machine-learning-repo-ubuntu1804_1.0.0-1_amd64.deb | |||
sudo apt install ./nvidia-machine-learning-repo-ubuntu1804_1.0.0-1_amd64.deb | |||
sudo apt-get update | |||
# Install NVIDIA driver | |||
sudo apt-get install --no-install-recommends nvidia-driver-430 | |||
# Reboot. Check that GPUs are visible using the command: nvidia-smi | |||
# Install development and runtime libraries (~4GB) | |||
sudo apt-get install --no-install-recommends \ | |||
cuda-10-1 \ | |||
libcudnn7=7.6.4.38-1+cuda10.1 \ | |||
libcudnn7-dev=7.6.4.38-1+cuda10.1 | |||
# Install TensorRT. Requires that libcudnn7 is installed above. | |||
sudo apt-get install -y --no-install-recommends libnvinfer6=6.0.1-1+cuda10.1 \ | |||
libnvinfer-dev=6.0.1-1+cuda10.1 \ | |||
libnvinfer-plugin6=6.0.1-1+cuda10.1 | |||
</pre> | |||
Revision as of 19:03, 8 June 2020
Installation
Linux
- Install the latest nvidia drivers from the standard repo, e.g.
nvidia-drivers-440
- Install Cuda Toolkit separately without the drivers.
- Use one of the deb install options.
- You may also want to install the following:
- cuDnn
- TensorRT
- cuDnn
- Copied from tensorflow
# Add NVIDIA package repositories wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/cuda-repo-ubuntu1804_10.1.243-1_amd64.deb sudo apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/7fa2af80.pub sudo dpkg -i cuda-repo-ubuntu1804_10.1.243-1_amd64.deb sudo apt-get update wget http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1804/x86_64/nvidia-machine-learning-repo-ubuntu1804_1.0.0-1_amd64.deb sudo apt install ./nvidia-machine-learning-repo-ubuntu1804_1.0.0-1_amd64.deb sudo apt-get update # Install NVIDIA driver sudo apt-get install --no-install-recommends nvidia-driver-430 # Reboot. Check that GPUs are visible using the command: nvidia-smi # Install development and runtime libraries (~4GB) sudo apt-get install --no-install-recommends \ cuda-10-1 \ libcudnn7=7.6.4.38-1+cuda10.1 \ libcudnn7-dev=7.6.4.38-1+cuda10.1 # Install TensorRT. Requires that libcudnn7 is installed above. sudo apt-get install -y --no-install-recommends libnvinfer6=6.0.1-1+cuda10.1 \ libnvinfer-dev=6.0.1-1+cuda10.1 \ libnvinfer-plugin6=6.0.1-1+cuda10.1
For tensorflow and pytorch, you may need to add LD_LIBRARY_PATH=/usr/local/cuda/lib64
to your environment variables.
You can also do this in PyCharm.