NVIDIA Driver

https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/
Admin
NVIDIA Driver is a software package that enables GPU support on Linux systems, providing...
Tokens:30,748
Snippets:379
Trust Score:10
Update:2 months ago
Show doc for...
Tokens:
Raw
### Install NVIDIA Driver Packages

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/red-hat-enterprise-linux.html

Examples of installing various NVIDIA driver combinations using DNF.

```bash
# dnf install nvidia-open
# dnf install nvidia-driver-cuda
# dnf install nvidia-driver nvidia-settings
# dnf install nvidia-driver nvidia-driver-cuda
```

--------------------------------

### Switching Driver Branches via APT

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/version-locking.html

Example command and output for installing a new pinning package to switch driver branches.

```bash
# apt install -V --purge nvidia-driver-pinning-590
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
The following packages will be REMOVED:
   nvidia-driver-570 (570-0ubuntu1)
The following NEW packages will be installed:
   nvidia-driver-590 (590-0ubuntu1)
0 upgraded, 1 newly installed, 0 downgraded, 1 to remove and 0 not upgraded.
Need to get 10 KB of archives.
After this operation, 10 KB of additional disk space will be used.
Do you want to continue? [Y/n]
```

--------------------------------

### Install recommended driver

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/driver-assistant.html

Executes the assistant with the install flag to automatically proceed with the recommended driver installation.

```bash
$ nvidia-driver-assistant --install
```

--------------------------------

### NVIDIA Driver Pinning Package Examples

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/version-locking.html

Examples of branch-specific and version-specific pinning packages available for installation.

```text
nvidia-driver-pinning-570
nvidia-driver-pinning-575
nvidia-driver-pinning-580
nvidia-driver-pinning-590
```

```text
nvidia-driver-pinning-570.199
nvidia-driver-pinning-570.200
nvidia-driver-pinning-580.102.03
nvidia-driver-pinning-580.103
```

--------------------------------

### Install Local Repository

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/azure-linux.html

Install the downloaded local repository package on the file system.

```bash
# rpm --install nvidia-driver-local-repo-${distro}.${version}*.${arch}.rpm
```

--------------------------------

### Install Standard Drivers

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/fedora.html

Install either open or proprietary kernel modules for standard driver setups.

```bash
# dnf install nvidia-open
```

```bash
# dnf install cuda-drivers
```

--------------------------------

### Enable Installation Logging

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/windows.html

Add log directory and level parameters to the silent installation command to capture installation details.

```bash
setup.exe -s -n Display.Driver -log:c:\logs -loglevel:6
```

--------------------------------

### Install NVSwitch components on SUSE/openSUSE

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/optional-components.html

Use zypper to install the required packages on SUSE-based systems.

```bash
# zypper install nvidia-fabricmanager libnvidia-nscq libnvsdm nvidia-imex
```

--------------------------------

### Install NVSwitch components on Ubuntu/Debian

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/optional-components.html

Use apt to install the required packages on Debian-based systems.

```bash
# apt install -V nvidia-fabricmanager libnvidia-nscq libnvsdm nvidia-imex
```

--------------------------------

### Install Network Repository via cuda-keyring

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/debian.html

Install the cuda-keyring package to enable the network repository.

```bash
$ wget https://developer.download.nvidia.com/compute/cuda/repos/${distro}/x86_64/cuda-keyring_1.1-1_all.deb
# dpkg -i cuda-keyring_1.1-1_all.deb
# apt update
```

--------------------------------

### Install 32-bit libraries on SUSE/OpenSUSE

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/optional-components.html

Installs 32-bit G07 NVIDIA compute, GL, and video libraries using zypper.

```bash
# zypper install \
   nvidia-compute-G07-32bit \
   nvidia-gl-G07-32bit \
   nvidia-video-G07-32bit
```

--------------------------------

### Install NVIDIA Driver

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/azure-linux.html

Install the NVIDIA open-source driver package.

```bash
# tdnf install nvidia-open
```

--------------------------------

### Install Open Kernel Modules

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/amazon-linux.html

Enable and install the open-source NVIDIA kernel modules.

```bash
# dnf module enable nvidia-driver:open-dkms
# dnf install nvidia-open
```

--------------------------------

### Install 32-bit libraries on RHEL/Alma/Rocky/Oracle/Fedora

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/optional-components.html

Installs 32-bit NVIDIA driver and FBC libraries using dnf.

```bash
# dnf install \
   nvidia-driver-cuda-libs.i686 \
   nvidia-driver-libs.i686 \
   libnvidia-fbc.i686
```

--------------------------------

### Install Kernel Headers

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/amazon-linux.html

Install the kernel headers and development packages for the currently running kernel.

```bash
# dnf install kernel-devel-$(uname -r) kernel-headers-$(uname -r)
```

--------------------------------

### Install Local Repository

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/debian.html

Install the downloaded local repository package and update the apt cache.

```bash
# dpkg -i nvidia-driver-local-repo-${distro}-${version}*_${arch}.deb
# apt update
```

--------------------------------

### Download and Install Local Repository

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/fedora.html

Commands to download the NVIDIA driver repository RPM and install it on the local file system.

```bash
$ wget https://developer.download.nvidia.com/compute/nvidia-driver/${version}/local_installers/nvidia-driver-local-repo-${distro}.${version}.${arch}.rpm
```

```bash
# rpm --install nvidia-driver-local-repo-${distro}.${version}*.${arch}.rpm
```

--------------------------------

### Install rebuilt nvidia-peermem

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/optional-components.html

Install the newly built modules back into the system kernel directory.

```bash
# dkms install -m nvidia/590.44.01
Installing /lib/modules/6.17.11-300.fc42.x86_64/extra/nvidia.ko.xz
Installing /lib/modules/6.17.11-300.fc42.x86_64/extra/nvidia-modeset.ko.xz
Installing /lib/modules/6.17.11-300.fc42.x86_64/extra/nvidia-drm.ko.xz
Installing /lib/modules/6.17.11-300.fc42.x86_64/extra/nvidia-uvm.ko.xz
Installing /lib/modules/6.17.11-300.fc42.x86_64/extra/nvidia-peermem.ko.xz
Running depmod...... done.
Executing post-transaction command.......... done.
```

--------------------------------

### Install Desktop-only Drivers

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/fedora.html

Install drivers for desktop systems without compute components.

```bash
# dnf install nvidia-driver kmod-nvidia-open-dkms
```

```bash
# dnf install nvidia-driver kmod-nvidia-latest-dkms
```

--------------------------------

### Azure Linux dependency resolution example

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/optional-components.html

Example demonstrating potential dependency resolution issues when using tdnf.

```bash
root [ / ]# tdnf list nvidia-fabricmanager nvidia-fabric-manager
Loaded plugin: tdnfrepogpgcheck
nvidia-fabricmanager.x86_64          580.105.08-1        cuda-azl3-x86_64
nvidia-fabricmanager.x86_64          580.65.06-1         cuda-azl3-x86_64
nvidia-fabricmanager.x86_64          580.82.07-1         cuda-azl3-x86_64
nvidia-fabricmanager.x86_64          580.95.05-1         cuda-azl3-x86_64
nvidia-fabricmanager.x86_64          590.44.01-1.azl3    cuda-azl3-x86_64
nvidia-fabricmanager.x86_64          590.48.01-1.azl3    cuda-azl3-x86_64
nvidia-fabric-manager.x86_64         575.51.03-1         cuda-azl3-x86_64
nvidia-fabric-manager.x86_64         575.57.08-1         cuda-azl3-x86_64
root [ / ]# tdnf install nvidia-fabricmanager
Loaded plugin: tdnfrepogpgcheck

Installing:
nvidia-fabric-manager x86_64         575.57.08-1         cuda-azl3-x86_64  36.56M      10.06M

Total installed size:  36.56M
Total download size:  10.06M
Is this ok [y/N]: ^C
root [ / ]# tdnf install nvidia-fabricmanager --best
Loaded plugin: tdnfrepogpgcheck

Installing:
nvidia-fabric-manager x86_64         575.57.08-1         cuda-azl3-x86_64  36.56M      10.06M

Total installed size:  36.56M
Total download size:  10.06M
Is this ok [y/N]: ^C
```

--------------------------------

### Install NVIDIA Drivers

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/almalinux.html

Installs the NVIDIA open kernel modules and driver components.

```bash
# dnf install nvidia-open-kmod nvidia-driver nvidia-driver-cuda
```

```bash
# dnf install nvidia-open-kmod nvidia-driver nvidia-driver-cuda nvidia-driver-libs.i686
```

--------------------------------

### Run nvidia-driver-assistant

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/driver-assistant.html

Executes the assistant to detect system hardware and display the recommended installation command.

```bash
$ nvidia-driver-assistant
```

--------------------------------

### Install Proprietary Kernel Modules

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/amazon-linux.html

Enable and install the proprietary NVIDIA kernel modules.

```bash
# dnf module enable nvidia-driver:latest-dkms
# dnf install cuda-drivers
```

--------------------------------

### Install Module Streams

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/amazon-linux.html

Install specific NVIDIA driver module streams.

```bash
# dnf module install nvidia-driver:<stream>/<profile>
```

--------------------------------

### Install NVSwitch components on RHEL/Alma/Rocky/Oracle/Amazon/Kylin

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/optional-components.html

Use the dnf module command to install NVSwitch components on supported RHEL-based distributions.

```bash
# dnf module install nvidia-driver:$stream/fm
```

--------------------------------

### Install Modularity Profiles via DNF

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/advanced-options.html

Use these commands to install specific driver profiles from a modularity stream.

```bash
# dnf module install nvidia-driver:<stream>/default
# dnf module install nvidia-driver:<stream>/fm
```

--------------------------------

### Install 32-bit libraries on Ubuntu 22.04/24.04/26.04

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/optional-components.html

Enables i386 architecture and installs specific 32-bit NVIDIA compute, decode, encode, FBC, and GL libraries.

```bash
# dpkg --add-architecture i386
# apt update
# apt install \
   libnvidia-compute:i386 \
   libnvidia-decode:i386 \
   libnvidia-encode:i386 \
   libnvidia-fbc1:i386 \
   libnvidia-gl:i386
```

--------------------------------

### Install Compute-only Drivers

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/fedora.html

Install drivers for headless systems without desktop components.

```bash
# dnf install nvidia-driver-cuda kmod-nvidia-open-dkms
```

```bash
# dnf install nvidia-driver-cuda kmod-nvidia-latest-dkms
```

--------------------------------

### Enable Network Repository

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/azure-linux.html

Configure the network repository for driver installation based on architecture.

```bash
# wget https://developer.download.nvidia.com/compute/cuda/repos/${distro}/${arch}/cuda-${distro}.repo -O /etc/yum.repos.d/cuda-${distro}.repo
```

```bash
# wget https://developer.download.nvidia.com/compute/cuda/repos/${distro}/sbsa/cuda-${distro}.repo -O /etc/yum.repos.d/cuda-${distro}.repo
```

--------------------------------

### Enable Network Repository via cuda-keyring

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/debian.html

Install the cuda-keyring package to configure the repository for arm64 systems.

```bash
$ wget https://developer.download.nvidia.com/compute/cuda/repos/${distro}/sbsa/cuda-keyring_1.1-1_all.deb
# dpkg -i cuda-keyring_1.1-1_all.deb
# apt update
```

--------------------------------

### Configure GPU preference in desktop files

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/optimus-laptops-and-multi-gpu-desktop-systems.html

Example of desktop entry settings to force an application to run on a discrete GPU.

```bash
$ cat /usr/share/applications/steam.desktop | grep -i GPU
PrefersNonDefaultGPU=true
X-KDE-RunOnDiscreteGpu=true
```

--------------------------------

### Install NVIDIA Drivers

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/ubuntu.html

Commands to install either open or proprietary kernel modules on Ubuntu.

```bash
# apt install nvidia-open
```

```bash
# apt install cuda-drivers
```

--------------------------------

### Install Kernel Dependencies

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/azure-linux.html

Install the necessary kernel headers and development packages for the currently running kernel.

```bash
# tdnf install kernel-devel-$(uname -r) kernel-headers-$(uname -r) kernel-modules-extra-$(uname -r)
```

--------------------------------

### Install NVSwitch components on RHEL 10/Alma 10/Rocky 10/Fedora 44

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/optional-components.html

Use dnf to install the required packages on newer RHEL-based systems.

```bash
# dnf install nvidia-fabricmanager libnvidia-nscq libnvsdm nvidia-imex
```

--------------------------------

### Install GPUDirect Storage on RHEL 8/9

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/optional-components.html

Installs the nvidia-fs package for GPUDirect Storage support.

```bash
# dnf install nvidia-fs
```

--------------------------------

### Manual Repository Signing Key Enrollment

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/debian.html

Use this method if the cuda-keyring package installation fails.

```bash
$ wget https://developer.download.nvidia.com/compute/cuda/repos/${distro}/sbsa/cuda-archive-keyring.gpg
# mv cuda-archive-keyring.gpg /usr/share/keyrings/cuda-archive-keyring.gpg
```

--------------------------------

### Install NVSwitch components on Azure Linux

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/optional-components.html

Use tdnf to install the required packages on Azure Linux.

```bash
# tdnf install nvidia-fabricmanager libnvidia-nscq libnvsdm nvidia-imex
```

--------------------------------

### Install kernel development packages

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/suse.html

Installs the necessary kernel development packages for the current system.

```bash
# zypper install -y kernel-<variant>-devel=<version>
```

--------------------------------

### Install NVIDIA Kernel Modules

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/debian.html

Install either open or proprietary kernel modules for the driver.

```bash
# apt -V install nvidia-open
```

```bash
# apt -V install cuda-drivers
```

--------------------------------

### Install Packages with Verbose Versions

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/advanced-options.html

Use the --verbose-versions flag to display detailed version information during package installation.

```bash
# apt install --verbose-versions nvidia-open
```

--------------------------------

### Install 32-bit libraries on Debian 12/13

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/optional-components.html

Enables i386 architecture and installs the necessary 32-bit NVIDIA driver libraries.

```bash
# dpkg --add-architecture i386
# apt update
# apt install nvidia-driver-libs:i386
```

--------------------------------

### Install layer library on Debian/Ubuntu

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/gaming.html

Install the shared object file to the appropriate library directory for Debian-based distributions.

```bash
# install -m0755 -D libdxvk_nvapi_vkreflex_layer.so \
    /usr/local/lib/x86_64-linux-gnu/libdxvk_nvapi_vkreflex_layer.so
```

--------------------------------

### Install kernel-default-devel package

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/suse.html

Installs the kernel development package using the current kernel version.

```bash
# zypper install -y kernel-default-devel=$(uname -r | sed 's/\-default//')
```

--------------------------------

### Install layer library on RPM distributions

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/gaming.html

Install the shared object file to the appropriate library directory for RPM-based distributions.

```bash
# install -m0755 -D libdxvk_nvapi_vkreflex_layer.so \
   /usr/local/lib64/libdxvk_nvapi_vkreflex_layer.so
```

--------------------------------

### Install Multiple Modularity Profiles

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/advanced-options.html

Utilize BASH curly brace expansion to install multiple profiles in a single command.

```bash
# dnf module install nvidia-driver:latest/{default,fm}
```

--------------------------------

### Install wine-nvml for Proton 10

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/gaming.html

Manually install the wine-nvml library to enable GPU utilization reporting in Proton 10 environments.

```bash
$ PROTON_LIBS=~/.local/share/Steam/steamapps/common/Proton\ 10.0/files/lib/wine
$ unzip wine-nvml*.zip
$ install -m0555 -p lib64/wine/x86_64-windows/nvml.dll "${PROTON_LIBS}/x86_64-windows/"
$ install -m0555 -p lib64/wine/x86_64-unix/nvml.so "${PROTON_LIBS}/x86_64-unix/"
```

--------------------------------

### Configure Repositories on AlmaLinux 8

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/almalinux.html

Enable the PowerTools repository and install the EPEL release package for dependency resolution.

```bash
# dnf config-manager --set-enabled powertools
# dnf install epel-release
```

--------------------------------

### Install GPUDirect Storage on Ubuntu/Debian

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/optional-components.html

Installs the nvidia-fs package for GPUDirect Storage support on Debian-based systems.

```bash
# apt install nvidia-fs
```

--------------------------------

### Switch Driver Modules on SUSE/OpenSUSE

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/advanced-options.html

Use zypper to install the target driver module with force resolution enabled.

```bash
# zypper install --details --force-resolution nvidia-open
```

```bash
# zypper install --details --force-resolution cuda-drivers
```

--------------------------------

### Install the Vulkan Reflex layer

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/gaming.html

Extract the layer files and update the JSON configuration path for the DXVK-NVAPI Reflex layer.

```bash
$ tar -xzf dxvk-nvapi-*.tar.gz --strip-components=2 ./layer
$ sed -i -e 's|./libdxvk|libdxvk|g' VkLayer_DXVK_NVAPI_reflex.json
# install -m0644 -D VkLayer_DXVK_NVAPI_reflex.json \
    /usr/local/share/vulkan/implicit_layer.d/VkLayer_DXVK_NVAPI_reflex.json
```

--------------------------------

### Install closed module flavor

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/driver-assistant.html

Executes the assistant to install the closed kernel module flavor specifically.

```bash
$ nvidia-driver-assistant --install --module-flavor closed
```

--------------------------------

### Switch Driver Modules on RHEL-based Systems (10/Fedora)

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/advanced-options.html

Use dnf to install the specific driver package to switch between proprietary and open modules.

```bash
# dnf install --allowerasing nvidia-open
```

```bash
# dnf install --allowerasing cuda-drivers
```

--------------------------------

### Configure Repositories on AlmaLinux 9/10

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/almalinux.html

Enable the CRB repository and install the EPEL release package for dependency resolution.

```bash
# dnf config-manager --set-enabled crb
# dnf install epel-release
```

--------------------------------

### Install NVIDIA kernel module with DKMS

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/frequently-asked-questions.html

Install the built NVIDIA kernel module into the system kernel directory.

```bash
# dkms -m nvidia/565.57.01 -k 6.11.7-300.fc41.x86_64 install

Module nvidia-565.57.01 for kernel 6.11.7-300.fc41.x86_64 (x86_64):
Before uninstall, this module version was ACTIVE on this kernel.
Deleting /lib/modules/6.11.7-300.fc41.x86_64/extra/nvidia.ko.xz
Deleting /lib/modules/6.11.7-300.fc41.x86_64/extra/nvidia-modeset.ko.xz
Deleting /lib/modules/6.11.7-300.fc41.x86_64/extra/nvidia-drm.ko.xz
Deleting /lib/modules/6.11.7-300.fc41.x86_64/extra/nvidia-uvm.ko.xz
Deleting /lib/modules/6.11.7-300.fc41.x86_64/extra/nvidia-peermem.ko.xz
Running depmod.... done.

Installing /lib/modules/6.11.7-300.fc41.x86_64/extra/nvidia.ko.xz
Installing /lib/modules/6.11.7-300.fc41.x86_64/extra/nvidia-modeset.ko.xz
Installing /lib/modules/6.11.7-300.fc41.x86_64/extra/nvidia-drm.ko.xz
Installing /lib/modules/6.11.7-300.fc41.x86_64/extra/nvidia-uvm.ko.xz
Installing /lib/modules/6.11.7-300.fc41.x86_64/extra/nvidia-peermem.ko.xz
Running depmod... done.

```

--------------------------------

### Configure Repositories for RHEL 10

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/red-hat-enterprise-linux.html

Enable required RHEL 10 repositories and install the EPEL release package.

```bash
# subscription-manager repos --enable=rhel-10-for-$arch-appstream-rpms
# subscription-manager repos --enable=rhel-10-for-$arch-baseos-rpms
# subscription-manager repos --enable=codeready-builder-for-rhel-10-$arch-rpms
# dnf install https://dl.fedoraproject.org/pub/epel/epel-release-latest-10.noarch.rpm
```

--------------------------------

### Configure Repositories for RHEL 9

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/red-hat-enterprise-linux.html

Enable required RHEL 9 repositories and install the EPEL release package.

```bash
# subscription-manager repos --enable=rhel-9-for-$arch-appstream-rpms
# subscription-manager repos --enable=rhel-9-for-$arch-baseos-rpms
# subscription-manager repos --enable=codeready-builder-for-rhel-9-$arch-rpms
# dnf install https://dl.fedoraproject.org/pub/epel/epel-release-latest-9.noarch.rpm
```

--------------------------------

### Install Desktop-only Drivers

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/debian.html

Commands to install drivers for desktop systems using either open or proprietary kernel modules.

```bash
# apt -V install nvidia-driver nvidia-kernel-open-dkms
```

```bash
# apt -V install nvidia-driver nvidia-kernel-dkms
```

--------------------------------

### Add network repository

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/suse.html

Adds the NVIDIA CUDA repository for network-based installation.

```bash
# zypper addrepo https://developer.download.nvidia.com/compute/cuda/repos/$distro/$arch/cuda-$distro.repo
```

```bash
# zypper addrepo https://developer.download.nvidia.com/compute/cuda/repos/$distro/sbsa/cuda-$distro.repo
```

--------------------------------

### Configure Repositories for RHEL 8

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/red-hat-enterprise-linux.html

Enable required RHEL 8 repositories and install the EPEL release package.

```bash
# subscription-manager repos --enable=rhel-8-for-$arch-appstream-rpms
# subscription-manager repos --enable=rhel-8-for-$arch-baseos-rpms
# subscription-manager repos --enable=codeready-builder-for-rhel-8-$arch-rpms
# dnf install https://dl.fedoraproject.org/pub/epel/epel-release-latest-8.noarch.rpm
```

--------------------------------

### Enable NVIDIA Driver Repositories

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/almalinux.html

Enables the necessary repositories for NVIDIA driver installation on AlmaLinux.

```bash
# dnf install almalinux-release-nvidia-driver
```

--------------------------------

### Install NVIDIA X.org driver on Ubuntu

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/wayland-only-desktop-installation.html

Explicitly installs the X.org component for the NVIDIA driver on Ubuntu systems where no reverse dependency exists.

```bash
# apt install xserver-xorg-video-nvidia
```

--------------------------------

### Perform Silent Driver Installation

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/windows.html

Use this command to install the display driver without the installer UI.

```bash
setup.exe -s -n Display.Driver
```

--------------------------------

### Enable network repository

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/ubuntu.html

Create the source list file to enable the network repository.

```bash
# echo "deb [signed-by=/usr/share/keyrings/cuda-archive-keyring.gpg] https://developer.download.nvidia.com/compute/cuda/repos/$distro/x86_64/ /" \
    | tee /etc/apt/sources.list.d/cuda-$distro-$arch.list
```

--------------------------------

### Switch Driver Modules on Debian

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/advanced-options.html

Use apt to remove existing drivers and install the target module flavor.

```bash
# apt install --autoremove --purge nvidia-open
```

```bash
# apt install --autoremove --purge cuda-drivers
```

--------------------------------

### Verify Vulkan layer installation

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/gaming.html

Check if the Vulkan layer is correctly recognized by the system using vulkaninfo.

```bash
$ vulkaninfo --summary | grep NVAPI
VK_LAYER_DXVK_NVAPI_reflex         DXVK-NVAPI Vulkan Reflex compatibility layer   1.3.295  version 1
```

--------------------------------

### Enable Network Repository Manually

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/debian.html

Create the source list file to enable the network repository with the specified signing key.

```bash
# echo "deb [signed-by=/usr/share/keyrings/cuda-archive-keyring.gpg] https://developer.download.nvidia.com/compute/cuda/repos/${distro}/x86_64/ /" \
    | tee /etc/apt/sources.list.d/cuda-${distro}-${arch}.list
```

--------------------------------

### Remove Local Repository Installer

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/post-installation-actions.html

Remove the local repository installer packages after the driver installation is complete, using the appropriate package manager for the distribution.

```bash
# dnf remove nvidia-driver-local-repo\*
```

```bash
# tdnf remove nvidia-driver-local-repo-\*
```

```bash
# apt remove --purge nvidia-driver-local-repo\*
```

```bash
# zypper remove nvidia-driver-local-repo\*
```

--------------------------------

### Enroll Signing Key Manually

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/debian.html

Manually download and move the signing key to the system keyrings directory if the keyring package installation fails.

```bash
$ wget https://developer.download.nvidia.com/compute/cuda/repos/${distro}/x86_64/cuda-archive-keyring.gpg
# mv cuda-archive-keyring.gpg /usr/share/keyrings/cuda-archive-keyring.gpg
```

--------------------------------

### Switching NVIDIA Driver Versions on Debian

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/version-locking.html

Use apt-get to install a specific pinning package and perform a distribution upgrade to apply the version change.

```bash
# apt-get install nvidia-driver-pinning-570.199
[..]
# apt-get -V dist-upgrade --autoremove
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
Calculating upgrade... Done
The following packages will be REMOVED:
   libnvidia-present (580.103-1)
The following packages will be DOWNGRADED:
   firmware-nvidia-gsp (580.103-1 => 570.199-1)
   libcuda1 (580.103-1 => 570.199-1)
   libcuda1:i386 (580.103-1 => 570.199-1)
   libcudadebugger1 (580.103-1 => 570.199-1)
   libegl-nvidia0 (580.103-1 => 570.199-1)
   libegl-nvidia0:i386 (580.103-1 => 570.199-1)
   libgles-nvidia1 (580.103-1 => 570.199-1)
   libgles-nvidia1:i386 (580.103-1 => 570.199-1)
   libgles-nvidia2 (580.103-1 => 570.199-1)
   libgles-nvidia2:i386 (580.103-1 => 570.199-1)
   libglx-nvidia0 (580.103-1 => 570.199-1)
   libglx-nvidia0:i386 (580.103-1 => 570.199-1)
   libnvcuvid1 (580.103-1 => 570.199-1)
   libnvidia-allocator1 (580.103-1 => 570.199-1)
   libnvidia-allocator1:i386 (580.103-1 => 570.199-1)
   libnvidia-api1 (580.103-1 => 570.199-1)
   libnvidia-cfg1 (580.103-1 => 570.199-1)
   libnvidia-eglcore (580.103-1 => 570.199-1)
   libnvidia-eglcore:i386 (580.103-1 => 570.199-1)
   libnvidia-encode1 (580.103-1 => 570.199-1)
   libnvidia-fbc1 (580.103-1 => 570.199-1)
   libnvidia-glcore (580.103-1 => 570.199-1)
   libnvidia-glcore:i386 (580.103-1 => 570.199-1)
   libnvidia-glvkspirv (580.103-1 => 570.199-1)
   libnvidia-glvkspirv:i386 (580.103-1 => 570.199-1)
   libnvidia-gpucomp (580.103-1 => 570.199-1)
   libnvidia-gpucomp:i386 (580.103-1 => 570.199-1)
   libnvidia-ml1 (580.103-1 => 570.199-1)
   libnvidia-ml1:i386 (580.103-1 => 570.199-1)
   libnvidia-ngx1 (580.103-1 => 570.199-1)
   libnvidia-nvvm4 (580.103-1 => 570.199-1)
   libnvidia-opticalflow1 (580.103-1 => 570.199-1)
   libnvidia-pkcs11-openssl3 (580.103-1 => 570.199-1)
   libnvidia-ptxjitcompiler1 (580.103-1 => 570.199-1)
   libnvidia-ptxjitcompiler1:i386 (580.103-1 => 570.199-1)
   libnvidia-rtcore (580.103-1 => 570.199-1)
   libnvidia-sandboxutils (580.103-1 => 570.199-1)
   libnvidia-vksc-core (580.103-1 => 570.199-1)
   libnvoptix1 (580.103-1 => 570.199-1)
   libxnvctrl0 (580.103-1 => 570.199-1)
   nvidia-driver (580.103-1 => 570.199-1)
   nvidia-driver-cuda (580.103-1 => 570.199-1)
   nvidia-driver-libs (580.103-1 => 570.199-1)
   nvidia-driver-libs:i386 (580.103-1 => 570.199-1)
   nvidia-egl-icd (580.103-1 => 570.199-1)
   nvidia-kernel-open-dkms (580.103-1 => 570.199-1)
   nvidia-kernel-support (580.103-1 => 570.199-1)
   nvidia-modprobe (580.103-1 => 570.199-1)
   nvidia-open (580.103-1 => 570.199-1)
   nvidia-opencl-icd (580.103-1 => 570.199-1)
   nvidia-persistenced (580.103-1 => 570.199-1)
   nvidia-settings (580.103-1 => 570.199-1)
   nvidia-vdpau-driver (580.103-1 => 570.199-1)
   nvidia-vulkan-icd (580.103-1 => 570.199-1)
   nvidia-vulkan-icd:i386 (580.103-1 => 570.199-1)
   nvidia-xconfig (580.103-1 => 570.199-1)
   xserver-xorg-video-nvidia (580.103-1 => 570.199-1)
0 upgraded, 0 newly installed, 57 downgraded, 1 to remove and 0 not upgraded.
Need to get 0 B/341 MB of archives.
After this operation, 171 MB disk space will be freed.
Do you want to continue? [Y/n]
```

--------------------------------

### Install kernel development packages for 64k kernel on aarch64

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/oracle-linux.html

Use this command when utilizing the 64k kernel variant on aarch64 systems.

```bash
# dnf install kernel-64k-devel-matched kernel-headers
```

--------------------------------

### Enable Network Repository Manually

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/debian.html

Configure the apt source list file manually after enrolling the signing key.

```bash
# echo "deb [signed-by=/usr/share/keyrings/cuda-archive-keyring.gpg] https://developer.download.nvidia.com/compute/cuda/repos/${distro}/sbsa/ /" \
    | tee /etc/apt/sources.list.d/cuda-${distro}-${arch}.list
```

--------------------------------

### Switch to Open Modules

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/amazon-linux.html

Switch from proprietary modules to open modules.

```bash
# dnf module reset nvidia-driver
# dnf module enable nvidia-driver:<stream>
# dnf install nvidia-open --allowerasing
```

--------------------------------

### Configure EPEL Repository for RHEL 9 and 10

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/advanced-options.html

Add the EPEL repository to the Kickstart configuration for RHEL 9, 10, and compatible distributions.

```text
repo --name=epel --baseurl=http://download.fedoraproject.org/pub/epel/9/Everything/${arch}
```

--------------------------------

### Update Bootloader in Kickstart

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/advanced-options.html

Add this command to the %post section to update bootloader parameters like disabling nouveau.

```text
nvidia-boot-update post
```

--------------------------------

### Configure EPEL Repository for RHEL 8

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/advanced-options.html

Add the EPEL repository to the Kickstart configuration for RHEL 8 and compatible distributions.

```text
repo --name=epel --baseurl=http://download.fedoraproject.org/pub/epel/8/Everything/${arch}
```

--------------------------------

### Install specific major version on Azure Linux

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/optional-components.html

Workaround for dependency issues by specifying the major version during installation.

```bash
# tdnf install nvidia-fabricmanager-590
```

--------------------------------

### List available GPUs with switcherooctl

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/optimus-laptops-and-multi-gpu-desktop-systems.html

Commands to display GPU devices and their associated environment variables for Prime/VGA Switcheroo.

```bash
$ switcherooctl list
Device: 0
  Name:        Intel Corporation Raptor Lake-P [Iris Xe Graphics]
  Default:     yes
  Environment: DRI_PRIME=pci-0000_00_02_0

Device: 1
  Name:        NVIDIA Corporation AD104GLM [RTX 3500 Ada Generation Laptop GPU]
  Default:     no
  Environment: DRI_PRIME=pci-0000_01_00_0
```

```bash
$ switcherooctl list
Device: 0
  Name:        Intel Corporation Raptor Lake-P [Iris Xe Graphics]
  Default:     yes
  Environment: DRI_PRIME=pci-0000_00_02_0

Device: 1
  Name:        NVIDIA Corporation AD104GLM [RTX 3500 Ada Generation Laptop GPU]
  Default:     no
  Environment: __GLX_VENDOR_LIBRARY_NAME=nvidia __NV_PRIME_RENDER_OFFLOAD=1 __VK_LAYER_NV_optimus=NVIDIA_only
```

--------------------------------

### Configure CUDA Repository for KylinOS 11

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/advanced-options.html

Add the CUDA repository to the Kickstart configuration for KylinOS 11.

```text
repo --name=cuda-rhel9 --baseurl=https://developer.download.nvidia.com/compute/cuda/repos/kylin11/${arch}/
```

--------------------------------

### Configure APT Preferences for NVIDIA Drivers

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/version-locking.html

Create or update /etc/apt/preferences.d/nvidia to pin specific driver versions. A priority of 1000 or higher allows for potential downgrades.

```text
# Basic driver packages, includes foreign architectures
Package: src:nvidia-graphics-drivers-580:any src:nvidia-kmod-open:any
Pin: version 580*
Pin-Priority: 1000

# Basic driver packages, only in the native architectures
Package: src:nvidia-modprobe src:nvidia-modprobe src:nvidia-persistenced src:nvidia-settings src:nvidia-xconfig
Pin: version 580*
Pin-Priority: 1000

# Meta packages, includes foreign architectures
Package: src:cuda-drivers:any src:nvidia-open:any
Pin: version 580*
Pin-Priority: 1000

# Extra driver packages, only in the native architectures
Package: src:libnvidia-nscq src:libnvsdm src:nvidia-fabricmanager src:nvidia-imex src:nvlink5 src:fabricmanager src:imex
Pin: version 580*
Pin-Priority: 1000
```

```text
# Basic driver packages, includes foreign architectures
Package: src:nvidia-graphics-drivers:any src:nvidia-kmod-open:any
Pin: version 580*
Pin-Priority: 1000

# Basic driver packages, only in the native architectures
Package: src:nvidia-modprobe src:nvidia-modprobe src:nvidia-persistenced src:nvidia-settings src:nvidia-xconfig
Pin: version 580*
Pin-Priority: 1000

# Meta packages, includes foreign architectures
Package: src:cuda-drivers:any src:nvidia-open:any
Pin: version 580*
Pin-Priority: 1000

# Extra driver packages, only in the native architectures
Package: src:libnvidia-nscq src:libnvsdm src:nvidia-fabricmanager src:nvidia-imex src:nvlink5 src:fabricmanager src:imex
Pin: version 580*
Pin-Priority: 1000
```

--------------------------------

### Enable Contrib Repository

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/debian.html

Enable the contrib repository for the system.

```bash
# add-apt-repository contrib
```

--------------------------------

### Enable EPEL and CodeReady Builder on Oracle Linux 8

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/oracle-linux.html

Configure the necessary repositories for Oracle Linux 8 to satisfy driver dependencies.

```bash
# dnf config-manager --set-enabled ol8_codeready_builder
# dnf install oracle-epel-release-el8
```

--------------------------------

### Download Local Repository

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/azure-linux.html

Download the NVIDIA driver local repository package.

```bash
$ wget https://developer.download.nvidia.com/compute/nvidia-driver/${version}/local_installers/nvidia-driver-local-repo-${distro}.${version}.${arch}.rpm
```

--------------------------------

### Install precompiled module source on RHEL-based systems

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/optional-components.html

Installs additional packages containing the source of precompiled modules for driver versions up to 565.

```bash
# dnf module install nvidia-driver:$stream/src
```

--------------------------------

### Enable EPEL and CodeReady Builder on Oracle Linux 9

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/oracle-linux.html

Configure the necessary repositories for Oracle Linux 9 to satisfy driver dependencies.

```bash
# dnf config-manager --set-enabled ol9_codeready_builder
# dnf install oracle-epel-release-el9
```

--------------------------------

### Install kernel development packages for UEK kernel

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/oracle-linux.html

Install the required kernel headers and development packages for UEK kernel variants on Oracle Linux 8 or 9.

```bash
# dnf install kernel-uek-devel-$(uname -r) kernel-headers
```

--------------------------------

### Install Proprietary GPU Kernel Modules

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/kernel-modules.html

Commands for installing proprietary kernel modules across various Linux distributions, including options for latest or specific branch versions.

```bash
# dnf module enable nvidia-driver:latest-dkms
# dnf install cuda-drivers
```

```bash
# dnf module enable nvidia-driver:580-dkms
# dnf install cuda-drivers
```

```bash
# dnf install cuda-drivers
```

```bash
# dnf versionlock \*nvidia\*580\*
# dnf install cuda-drivers
```

```bash
# dnf versionlock add \*nvidia\*580\*
# dnf install cuda-drivers
```

```bash
# zypper install cuda-drivers
```

```bash
# zypper addlock "*nvidia* >= 590"
# zypper install cuda-drivers
```

```bash
# apt install cuda-drivers
```

```bash
# apt install nvidia-driver-pinning-580
# apt install cuda-drivers
```

--------------------------------

### Install Open GPU Kernel Modules

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/kernel-modules.html

Commands for installing open-source kernel modules across various Linux distributions, including options for latest or specific branch versions.

```bash
# dnf module enable nvidia-driver:open-dkms
# dnf install nvidia-open
```

```bash
# dnf module enable nvidia-driver:580-open
# dnf install nvidia-open
```

```bash
# dnf install nvidia-open
```

```bash
# dnf versionlock \*nvidia\*580\*
# dnf install nvidia-open
```

```bash
# dnf versionlock add \*nvidia\*580\*
# dnf install nvidia-open
```

```bash
# zypper install nvidia-open
```

```bash
# zypper addlock "*nvidia* >= 590"
# zypper install nvidia-open
```

```bash
# apt install nvidia-open
```

```bash
# apt install nvidia-driver-pinning-580
# apt install nvidia-open
```

--------------------------------

### Switch Driver Modules on Ubuntu

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/advanced-options.html

Use apt to switch between proprietary and open driver packages on Ubuntu systems.

```bash
# apt install --autoremove --purge nvidia-open nvidia-driver-open
```

```bash
# apt install --autoremove --purge cuda-drivers nvidia-driver
```

--------------------------------

### Enable Extended Repository

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/azure-linux.html

Enable the extended repository required for driver installation.

```bash
# tdnf install azurelinux-repos-extended
```

--------------------------------

### Check system GCC version

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/oracle-linux.html

Verify the currently installed system compiler version.

```bash
$ gcc --version | grep GCC
gcc (GCC) 11.5.0 20240719 (Red Hat 11.5.0-5.0.1)
```

--------------------------------

### Import MOK Key for Secure Boot

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/red-hat-enterprise-linux.html

Imports the downloaded public key into the system firmware for Secure Boot verification.

```bash
# mokutil --import *-public_key.der
input password: <hidden>
input password again: <hidden>
# mokutil --list-new --short
d56f7e6ace NVIDIA
# systemctl reboot
```

--------------------------------

### Upgrade and Switch Drivers

Source: https://docs.nvidia.com/datacenter/tesla/driver-installation-guide/latest/fedora.html

Commands for updating installed packages or switching between driver types.

```bash
# dnf update
```

```bash
# dnf install nvidia-open --allowerasing
```
Tokens:
Raw