CUDA: Difference between revisions

Latest revision as of 08:46, 12 December 2024

Installation

I suggest using conda to install cuda for version control your project.

Note that nvidia-smi lists the maximum CUDA version supported by the GPU driver, not the installed version of CUDA.
You can have a different version of CUDA installed in each conda environment, independently of the version supported by the GPU driver.

Conda

See nvidia/cuda-toolkit and nvidia/cuda-libraries-dev

For example:

# Install the runtime only
conda install -c "nvidia/label/cuda-11.8.0" cuda-toolkit
# Install the runtime and the development tools
conda install -c "nvidia/label/cuda-11.8.0" cuda-toolkit cuda-libraries-dev cuda-nvcc

Ubuntu

CUDA Toolkit

# Install drivers
sudo apt install nvidia-driver-565-open

GCC Versions

nvcc sometimes only supports older gcc/g++ versions.
To make it use those by default, create the following symlinks:

sudo ln -s /usr/bin/gcc-6 /usr/local/cuda/bin/gcc
sudo ln -s /usr/bin/g++-6 /usr/local/cuda/bin/g++

Alternatively, you can use -ccbin and point to your gcc:

-ccbin /usr/local/cuda/bin/gcc

References

An Even Easier Introduction To Cuda

@@ Line 1: / Line 1: @@
 ==Installation==
-===Ubuntu===
+I suggest using conda to install cuda for version control your project.
-[https://developer.nvidia.com/cuda-toolkit CUDA Toolkit]
-{{hidden | Details |
+Note that <code>nvidia-smi</code> lists the maximum CUDA version supported by the GPU driver, not the installed version of CUDA.<br>
-See [https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#ubuntu-installation CUDA Ubuntu Installation]
+You can have a different version of CUDA installed in each conda environment, independently of the version supported by the GPU driver.
-<syntaxhighlight lang="bash">
-# Set UBUNTU_VERSION to 2004 or 2204
-UBUNTU_VERSION=$(lsb_release -sr | sed -e 's/\.//g')
-# Add NVIDIA package repositories
+===Conda===
-wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu${UBUNTU_VERSION}/x86_64/cuda-ubuntu${UBUNTU_VERSION}.pin
+See [https://anaconda.org/nvidia/cuda-toolkit nvidia/cuda-toolkit] and [https://anaconda.org/nvidia/cuda-libraries-dev nvidia/cuda-libraries-dev]
-sudo mv cuda-ubuntu${UBUNTU_VERSION}.pin /etc/apt/preferences.d/cuda-repository-pin-600
-sudo apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu${UBUNTU_VERSION}/x86_64/3bf863cc.pub
-sudo add-apt-repository "deb https://developer.download.nvidia.com/compute/cuda/repos/ubuntu${UBUNTU_VERSION}/x86_64/ /"
-# Install NVIDIA driver and cuda.
+For example:
-sudo apt install nvidia-driver-515 cuda
+<syntaxhighlight lang="bash">
-# Reboot and check that the drivers are working with nvidia-smi
+# Install the runtime only
-sudo reboot
+conda install -c "nvidia/label/cuda-11.8.0" cuda-toolkit
+# Install the runtime and the development tools
-# Install cudnn
+conda install -c "nvidia/label/cuda-11.8.0" cuda-toolkit cuda-libraries-dev cuda-nvcc
-sudo apt install libcudnn8 libcudnn8-dev
 </syntaxhighlight>
-;Notes
+===Ubuntu===
-* For machine learning, use Anaconda or Docker's CUDA since different versions of TensorFlow and PyTorch require different CUDA versions.
+[https://developer.nvidia.com/cuda-toolkit CUDA Toolkit]
-You may need to add <code>LD_LIBRARY_PATH&#x3D;/usr/local/cuda/lib64</code> to your environment variables.<br>
+<syntaxhighlight lang="bash">
-You can also do this in PyCharm.<br>
+# Install drivers
-[[File:Pycharm LD LIBRARY PATH config.png| 200x200px]]
+sudo apt install nvidia-driver-565-open
-[[File:Pycharm LD LIBRARY PATH console config.png| 200x200px]]
+</syntaxhighlight>
-}}
-===Conda===
-See [https://anaconda.org/nvidia/cuda-toolkit nvidia/cuda-toolkit] and [https://anaconda.org/nvidia/cuda-libraries-dev nvidia/cuda-libraries-dev]
-==GCC Versions==
+===GCC Versions===
 <code>nvcc</code> sometimes only supports older gcc/g++ versions.
 To make it use those by default, create the following symlinks: