The Nvidia repository has been updated; here is the table that lists the current versions:
| Operating system | CentOS / RHEL | Fedora | rawhide |
|---|---|---|---|
| Driver branch | Long Lived | Short Lived Long Lived | Short Lived Long Lived Beta |
| Video Codec SDK | Yes | Yes | Yes |
| Architectures: x86_64 aarch64 | Yes | Yes | Yes |
| Basic nvidia driver: nvidia-driver nvidia-driver-libs nvidia-libXNVCtrl nvidia-kmod-common | Yes | Yes | Yes |
| CUDA libraries and tools: libnvidia-ml nvidia-driver-cuda nvidia-driver-cuda-libs nvidia-persistenced | Yes | Yes | Yes |
| OpenGL Framebuffer Capture: libnvidia-fbc | Yes | Yes | Yes |
| Nvidia tools: nvidia-modprobe nvidia-settings nvidia-xconfig | Yes | Yes | Yes |
| Binary kernel modules (kABI): kmod-nvidia | Yes | No | No |
| DKMS kernel modules: dkms-nvidia | Yes | Yes | Yes |
| aKMOD kernel modules: akmod-nvidia | No | Yes | Yes |
| 32 bit compatibility on x86_64: libnvidia-ml nvidia-libXNVCtrl nvidia-driver-libs nvidia-driver-cuda-libs | Yes | Yes | Yes |
| VDPAU libraries | Yes | Yes | Yes |
| EGLStream-based Wayland external platform | Yes | Yes | Yes |
| GBM EGL external platform library | Yes | Yes | Yes |
Quite a few things have changed, the packages are going towards providing the complete packaged CUDA stack. So far, only the GPU deployment kit has been inclued; and the packages allow for parallel installation with the Nvidia CUDA repositories by osboleting/updating packages as required. Here is some details on the things that have been implemented.
Table of Contents
X.org configuration
- Starting from Fedora 21, all driver X.org configuration can be managed by simply adding/removing X.org configuration snippets in
/etc/X11/xorg.conf.d. - Use new OutputClass directive on Fedora 21 X.org server 1.16 (and later) to load the driver and do not rely on an edited
/etc/X11/xorg.conffile. This also removes editing of thexorg.conffile from the package scriptlets. - Add the
IgnoreABIdirective by default on Fedora rawhide builds.
Kernel modules
- Add a new UDev rule in
nvidia-driver-cudafor thenvidia-uvmmodule and make X.org NVIDIA Files section to be loaded latest in case there are other packages providing a custom Files section (thanks Jan P. Springer for spotting these). - The binary
nvidia-modprobeis now SETUID, but its package is no longer a mandatory requirement for the drivers, so it will not get installed by default. - Now that both Redhat Enterprise Linux 7 and CentOS 7 have been released, binary modules (kABI) are now provided for these distributions.
CUDA support
- Added the GPU Deployment kit to the repository. This is constructed with NVML (NVIDIA Management Library) included with the drivers plus headers, docs and samples from a separate tarball. The separate tarball is using a different version number than the drivers. This is packaged in the
nvidia-driver-NVMLandnvidia-driver-NVML-develpackages. Installing these, thegpu-deployment-kitdependency provided by the CUDA repositories is preserved. - Along with NVML, the
nvidia-healthmonpackage is provided to monitor TESLA GPU clusters.
Along this, there is the usual assortment of packages refinement (syntax, RPMLint, optimizations, etc.). For additional details, please see the Nvidia driver page.
If you would like to test the CUDA packages please contact me and I will point you to a repository hosting the CUDA packages.
It’s really so frustrating. I’m trying to get CUDA working on Fedora 21. I’ve tried at least a dozen work-arounds, with no success. Why can’t there be any easy, sure-fire way to use my Tesla 20m on my computer, that used to be able to use it.
Why has Fedora been abandoned?
Sorry but I’m not following. And Fedora 21 has not been abandoned. A few users are using these CUDA packages and they are happy with it. Also upstream, Nvidia releases CUDA installers with Fedora 21 support.