Specifically, this (3D capable) GPU, if it matters:
The GPU & drivers worked fine with my previous Ubuntu installation.
Below are the exact setup steps I followed, after updating all packages and installing the kernel headers, etc.
sudo dnf install fedora-workstation-repositories
sudo dnf config-manager --set-enabled rpmfusion-nonfree-nvidia-driver
sudo dnf repository-packages rpmfusion-nonfree-nvidia-driver info
sudo dnf install akmod-nvidia nvidia* xorg-x11-drv-nvidia*
# wait 5 minutes, this should show the version number
modinfo -F version nvidia
# reboot, should have nvidia tools installed now
This is a desktop computer.
I have another desktop computer with a 1080Ti and installed Fedora two weeks ago and those setup steps worked fine. A week ago I tried installing Fedora 32 on this computer and had the same NVIDIA driver problem. Using Fedora 33 instead didn’t change anything. I suspect this graphics card is deficient in some way and I don’t know what to do to proceed.
I think you are having issues with the newer kernels. The problem is that Nvidia is not supporting new kernels until the middle of November as far as I know. What a experienced Fedora user here in Brazil oriented me to do with my 1060 Nvidia card was to do download in Koji the following packages from kernel 5.8.15-301:
Then you install all of them with dnf install. Here a link where you can find these packages:
After you do that, you boot into the new installed kernel and install the kmod-nvidia package from Rpmfusion and then reboot again into that kernel and uses it until Nvidia fix things. If you have the akmod already installed, you should probably take it off to install the kmod… but I’m not so sure, I’m a simple user.
That’s it… remember that I’m not a experienced user, and I’m just trying to give an option so you can deal with this problem. Maybe you should backup your files to recover than easily in case you end up stuck in the CLI and you feel like you will have difficulties in a situation like this.
Observe as well that I’m considering that you Nvidia GPU is normal one (no optimus-prime mess) from 2012 or later.
Also remember that Koji is not exactly Fedora repositories, and you will be downloading a build done by a kernel mantainer in Fedora. As a simple user, i can’t tell you how much safe it is compared to the official repositories, but talking with more experienced users, i decided it was secure enought for me.
Probably these older kernels installed manually won’t receive security updates. In an ideal world Nvidia would not stop supporting newer kernels, and if they did so, Fedora could have an linux-lts package like Arch or Solus for example. Despite I heard that Solus (that focus more on beginners) didn’t even allowed the linux-current kernel to upgrade to non-nvidia supported kernels.
If we all have Fedora 33 installed and are uptodate, wouldn’t we all have the same kernel? Why would some Nvidia cards work and not others? Why does everything work fine on my other Fedora computer with a 1080 GPU?
This is the kernel version I am on:
$ uname -a
Linux main 5.8.16-300.fc33.x86_64 #1 SMP Mon Oct 19 13:18:33 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux
I’m on 5.8, so the kernel should be supported then. What else could be the problem?
The kernel module forcing part is now included in the post-installation hook already, and is no more required to be run explicitly and the wait is also optional as the modules would be load completely before poweroff and before boot.
(Find the reference here. Even the tool I wrote used to make users wait for this long until v0.2.5 and it was so inconvenient )
Thanks, and I appreciate that AMD and Intel release open-source drivers. I would go with that but I need the Nvidia card and drivers so I can use CUDA for ML research. I wish Nvidia would release open-source drivers also.
I’m happy to be a team player and help with the testing. To be clear on the steps here, I need to install the Python package in this repo and run the NVAutoInstFedora32 tool, even though I have Fedora33 installed. I should try each of the different “installation modes” and see if any of them work. Right? If there are errors or something doesn’t work, should I report back here or open an issue in that repo?
I just took a moment to validate my assumption that pytorch can only run on Nvidia graphics cards and it seems I might not have been correct. I actually don’t care at all about Nvidia cards, I just want pytorch to work.
No dice. I tried every installation mode, most of which said “nothing to do” because the packages were already installed. The -plcuda and --vidacc ones installed a bunch of stuff.
Are there other ideas I can try?
[ # ] NVIDIA AUTOINSTALLER FOR FEDORA 32 AND ABOVE
[ # ] CHECKING FOR GPU COMPATIBILITY...
[ ! ] Compatibility infomation was obtained
[ ✔ ] One or more active NVIDIA GPUs were detected
01:00.0 VGA compatible controller: NVIDIA Corporation GM206 [GeForce GTX 960] (rev a1)
[ ✔ ] An single dedicated GPU setup was detected
[ # ] GATHERING CURRENT HOST INFORMATION...
[ ! ] Host information was gathered
System: Linux v5.8.16-300.fc33.x86_64
Version: #1 SMP Mon Oct 19 13:18:33 UTC 2020
Distribution: Fedora x86_64
[ # ] CHECKING FOR HOST COMPATIBILITY...
[ ✔ ] Supported OS detected
This tool is expected to work correctly here
[ ✘ ] Leaving installer