So I’ve been trying to install the proprietary Nvidia drivers on my homelab so I can get my fine ass art generated using Automatic1111 & Stable diffusion. I installed the Nvidia 510 server drivers, everything seems fine, then when I reboot, nothing. WTF Nvidia, why you gotta break X? Why is x even needed on a server driver. What’s your problem Nvidia!
nvidia has always been hostile to open source, as far back as i can remember.
back when nvidia bought 3dfx they took down the source code for the open 3dfx drivers within days, if not on the same day. i remember because i had just gotten myself a sweet voodoo 5 some weeks before that, and the great linux support was the reason i chose it… of course the driver code survived elsewhere, but it told me all i needed to know about that company.
also: linus’ rant wasn’t just a fun stunt, it was necessary to get nvidia to properly cooperate with the open source community if they want to keep making money running linux on their hardware.
Takes about 8 hrs to setup properly. But once you do set your Nvidia card with Linux, you just never update your OS and cry to sleep every night.
Linux is their bread and butter when it comes to servers and machine learning, but that’s a specialized environment and they don’t really care about general desktop use on arbitrary distros. They care about big businesses with big support contracts. Nobody’s running Wayland on their supercomputer clusters.
I cannot wait until architecture-agnostic ML libraries are dominant and I can kiss CUDA goodbye for good. I swear, 90% of my tech problems over the past 5 years have boiled down to “Nvidia sucks”. I’ve changed distros three times hoping it would make things easier, and it never really does; it just creates exciting new problems to play whack-a-mole with. I currently have Ubuntu LTS working, and I’m hoping I never need to breathe on it again.
That said, there’s honestly some grass-is-greener syndrome going on here, because you know what sucks almost as much as using Nvidia on Linux? Using Nvidia on Windows.
I cannot wait until architecture-agnostic ML libraries are dominant and I can kiss CUDA goodbye for good
I really hope this happens. After being on Nvidia for over a decade (960 for 5 years and similar midrange cards before that), I finally went AMD at the end of last year. Then of course AI burst onto the scene this year, and I’ve not yet managed to get stable diffusion running to the point it’s made me wonder if I might have made a bad choice.
It’s possible to run stable diffusion on amd cards, it’s just a bit more tedious and a lot slower. I managed to get it working on my rx 6700 under arch linux just fine. Now that I’m on fedora, it doesn’t really want to work for some reason, but I’m sure that it can be fixed as well, I just didn’t spend enough time on it.
It just makes no sense to me though, how is it sustainable for nvidia to not have great Linux kernel support? Like, let the kernel maintainers do their job and reap the benefits. I’m guessing that nvidia sees enterprise support contracts as an essential revenue stream, but eventually even enterprises are going to go with hardware that Linus isn’t giving the finger to right? Am I crazy?
Nvidia doesn’t hate linux, it just don’t care and the linux community hates nvidia
amd didn’t care a few years ago, but their drivers are open, so the community can fix it even if the company don’t care(now amd care a lot more, so it’s better) nvidia is a closed source crap, and it don’t give a fuck too
I thought nvidis was opening their drivers some as of sometime last year. Still shit though.
It didnt open the parts of the driver that mattered
The kernel modules were opened, which is absolutely relevant, it’ll make the entire process easier for many. The thing is that it’s not the same drivers, this open variety is pretty bad still and the code is simply not good enough yet to be incorporated into the Linux kernel. They were working on it, but it’s gonna take time.
And they can’t get all those sweet sweet tracking data they get from Windows users
Companies love to use open source software to reduce their development costs. They hate to contribute back.
Fine ass art. You’re in the lemmynsfw AI porn sub for sure.
They certainly have plenty of that, there… a friend told me.
What i don’t get is how nvidia stock is exploding when using their hardware for AI is a nightmare on Linux. How are companies doing this? Are they just offering enterprise support to ibsiders or something?
For what it’s worth, NVIDIA’s failings on Linux tend to be mostly in the desktop experience. As a compute device driven by cuda and not responsible for the display buffer, they work plenty good. Enterprise will not be running hardware GUI or DEs on the machines that do the AI work, if at all.
Even the old 1060 in my truenas scale server, has worked absolutely flawlessly with my jellyfin server.
They don’t give a fuck about consumers these days and Linux being just a tiny fraction of the userbase, they give even less of a fuck.
Nvidia is a breeze on linux vs amd. cuda is the only thing meaningfully supported across Windows and Linux. I fought with my 6900xt for so long trying to get ROCm working that I eventually bought a used 1080ti just to do the AI/ML stuff I wanted to do. I threw that into a server and had everything up and running in literally 10 minutes (and 5 minutes was making proxmox pass the gpu through to the VM).
People want to bitch about nvidia, but their entire ecosystem is better than AMD. The documentation is better and the tooling is better. On paper AMD is competitive but in practice Nvidia has so much more going for it–especially if you are doing any sort of AI/ML.
There are some benefits to to amd on linux; its the reason I replaced my 3070ti for a 6900xt. But that experience taught me: 1. AMD isn’t as good on linux as people give it credit for 2. nvidia isn’t as bad on linux as people blame it for. You trade different issues. Eg. Lose nvenc and cant use amf unless you use the amdpro driver not the open source one. if you use the pro driver you immediately lose half the benefits of the open source driver which is probably why you switch to amd on linux to begin with. So if you game, you can’t stream with a decent encoder–so you have to play with settings and throw cpu horsepower at it.
But hey, my DE doesn’t stutter and I dont have to do kludgy workarounds to get some games to play.
If AMD was able to come to the bright side, so can Nvidia. There’s still hope, ye faithful!
I’m on the cusp off jumping to Arch. Before I do I’m replacing my rtx 3080 with an RX 6800 XT. They are close enough in performance and identical pricing on eBay.
I’ve done a bunch of testing and found great support for all my hardware except my Razer Ripsaw HDMI capture device, which I can replace with something supported. It is just the Nvidia bullshit holding me back.
When I built my pc, I made sure to get AMD because of the nvidia outcry from the linux community. Thank goodness I got a 6800xt. I haven’t had any problems with it. It worked straight out of the box.
While I’m using AMD, I have had no issues with Nvidia on Arch using X before I switched earlier this year. One just installs the nvidia or nvidia-dkms package. My main reasons to switch were I had a 1060 6GB and it was getting old, AMD had a better price and if I’m keeping this one as long as my last I wanted to be certain wayland support was good even though I don’t use it right now
Using X, but X is shitty since ages and Wayland is gaining neat features that I personally don’t give a damn to (see: tearing protocol)
I’m gonna be that person… I rarely, if ever have issues with nvidia on Linux. Used several 30xx series cards for gaming over the last couple of years and it’s been a great experience.
Is it my distro (Void)?. is it because I’m happy staying on X11? Is it just luck? Interested to hear people’s gripes
I’m gonna be that person…
Well, you are not alone. While I too would prefer not to use proprietary drivers, I have had no problems on any of my Nvidia machines as well. Ironically, despite the open source drivers, getting a 7900XTX card up and running was an issue for me for months till distros caught up (with newer kernels and mesa libs), while my 4090 installation was a breeze even on the day it was released.
A lot of problems people have with Nvidia GPUs seem to be installation related. I think that is because the installation tends to be distro-specific and people do not necessarily follow the correct procedure for their distro or try installing the drivers directly from the Nvidia site as they would on Windows. For example, Fedora requires you to add RPMFusion, Debian needs non-free to be added to sources, Linux Mint lets you install the proprietary drivers but only after the first boot, and so on. Pop OS! probably makes the process the easiest with their Nvidia-specific ISO.
Minimal issues here. Set up Arch, install nVidia, add build hooks before next kernel update, carry on.
There is definitely some substance behind the complaints, but I think they are overblown or just the typical linux-user-parroting something they heard other people say.
On PopOS my 3070ti was always stable. I ran into occasional stuttering in the DE, but the biggest thing was I had to manual compile shaders using some guys github repo to play Apex Legends without it being a stuttery mess. But like you said, Pop is on X11 so maybe that makes a difference?
I bought into the “if you are going to use linux, especially for gaming, you need an amd gpu.” So I bought a 6900xt. I’ve had as many issues with my 6900xt; they are just different types of issues. Nothing insurmountable but its not like its some panacea.
I have a 3080 and it runs fine with openSUSE Tumbleweed. On first boot you do need to add the nvidia repo and then install it which I guess could be problematic for new linux users, but it’s literally pasting 1 line into terminal and then clicking the driver in yast. Echoing what others have said, I’d prefer if nvidia was a little less hostile to open source but frankly the driver just works, and works well. The only thing I’ve used besides openSUSE lately is Pop_OS and I believe the nvidia driver was installed automatically. If someone is having trouble getting the driver installed that seems to be a failure of the distro, not the user. You should be able to depend on your distros packaging to take care of this stuff.
I’ve seen this photo a bunch of times. Who is this guy? And why is high flipping the bird?
Linus Torvalds, creator of the linux kernel, flips the bird to nvidia and they deserve it
It’s the creator of the linux kernel. In the video he quote says “fuck you nvidia” lol
Other people have already told you who Linus is but if you want some context for the image it was from a Q&A: https://youtu.be/dmfDaxYhi9I?t=2864
Go to 47:46 if the link starts you at the beginning of the video (he starts to do the hand gesture at 49:28).
Here is an alternative Piped link(s): https://piped.video/dmfDaxYhi9I?t=2864
Piped is a privacy-respecting open-source alternative frontend to YouTube.
I’m open-source, check me out at GitHub.
Cheers. 👍
Some nerd with an attitude problem and a severe lack of social graces.
Are we just going to post this meme forever and ever here?
Even when we all perish, this meme will remain.
You could just as yourself: has anything changed or is the situation still the same?
Yes. Whenever it’s fitting to say “fuck Nvidia!” So likely forever and ever and always.
b-but muh free karmuh
Laughs in amdgpu pro
535 drivers have been working fine for me in Ubuntu and Manjaro.
They love to publish drivers that worked with like 1 release of X 5 years ago when the card came out and never update them.
Except when they update them and it breaks X.
