nvidia driver版本

1. 遇到可以先卸载 nvidia-uninstall
2. 然后重新安装 NVIDIA-Linux-x86_64-460.91.03.run

   wget https://us.download.nvidia.com/XFree86/Linux-x86_64/460.91.03/NVIDIA-Linux-x86_64-460.91.03.run;bash NVIDIA-Linux-x86_64-460.91.03.run

3. cuda 安装 https://developer.nvidia.com/cuda-downloads，需要注意cuda和驱动对应关系
4. from scratch install cuda , https://sarus.readthedocs.io/en/stable/user/custom-cuda-images.html
5. cuda 镜像参考 https://gitlab.com/nvidia/container-images/cuda/-/tree/master/dist
6. https://pytorch.org/get-started/previous-versions/

cuda和driver对应关系

1. cuda一般是向后兼容驱动。 既低版本的cuda可以向后兼容高版本的驱动，可以理解为cuda是应用，兼容系统升级。 https://docs.nvidia.com/cuda/cuda-toolkit-release-notes/index.html
2. cudnn和cuda兼容关系 https://developer.nvidia.com/rdp/cudnn-archive
3. cuda历史版本 https://developer.nvidia.com/cuda-toolkit-archive
4. driver版本下载 https://www.nvidia.com/download/index.aspx
5. Documentation Archives :: NVIDIA Deep Learning TensorRT Documentation

安装docker

参考 https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/install-guide.html

curl https://get.docker.com | sh && sudo systemctl --now enable docker

distribution=$(. /etc/os-release;echo $ID$VERSION_ID) \
      && curl -fsSL https://nvidia.github.io/libnvidia-container/gpgkey | sudo gpg --dearmor -o /usr/share/keyrings/nvidia-container-toolkit-keyring.gpg \
      && curl -s -L https://nvidia.github.io/libnvidia-container/$distribution/libnvidia-container.list | \
            sed 's#deb https://#deb [signed-by=/usr/share/keyrings/nvidia-container-toolkit-keyring.gpg] https://#g' | \
            sudo tee /etc/apt/sources.list.d/nvidia-container-toolkit.list

apt update && sudo apt-get install -y nvidia-container-toolkit

nvidia-ctk runtime configure --runtime=docker&& systemctl restart docker