It used to be this way and was one of my biggest complaints. It’s no longer this way. Drivers for my Nvidia card works fine on my mint and arch setup.
the upsides of buying from a Company that donates to OSS projects rather then not donating and only maintains proprietary drivers.
IK broadcom also does this too,but broadcom do have drivers in Mesa only for the Raspberry PI.This kinda reminded me of a scene in That 70s Show where Red Forman strongly recommended to his son that he should only fit accessories compatible with his 1969 Oldsmobile car.
I am on Debian with Nvidia 550 drivers… yeah.
Silly question but why?
I’ve learned it best to use nvidia drivers with nvidia cards and the AMD drivers with the AMD cards. I recommend this for performance.
Thank you for posting this!! I can’t get an erection. I tried using an AMD driver the other day with my NVIDIA card and was stumped why my screen was blank. I’d give you gold if I could!
Absolutely crazy take, I bet no one has thought of this.
Intel drivers:
I need an intel driver to turn the fucking useless onboard graphics off. for debian. any tips?
You can’t. Some laptops have the igpu as the dedicated driver of the display and can’t do hardware mux. If your laptop doesn’t offer the option in UEFI, it probably doesn’t support it.
thank you for this info. what do you think would happen if I did the following , more or less with the info indicated in my post here?
https://gist.github.com/pjobson/9e5f7349cf4f28bc82f82ea980047778
https://lemmy.world/post/43248486/22203270
any tips?
My laptop has this option in the bios settings
Not sure if it’s only for laptops but you could check there.I checked, it’s the easiest option and isn’t on my stupid MSI motherboard
Leave it to budget boards to exclude every possible useful setting but keep “boot from lan”
when I bought this fucker for £800 in 2018, I assure you it did not feel budger.
HOW MUCH
I thought it was a budget $150 board…
The link is to a laptop.
Also, it’s MSI for “multiple serious issues”
That’s the thing with AMD drivers, they’re the damn near perfect software. Doing lots of stuff yet you’d never know it’s there. It stays nicely out of the user’s way, you don’t even have to think about installing them and shit just works
Then there are the Nvidia drivers
They are not perfect, but their developers – 1 or 2 actually allocated to work on in-kernel drivers, such as Mario Limonciello – almost are.
I used em dashes to avoid a comma party, I promise I am not a LLM bot
I used em dashes to avoid a comma party, I promise I am not a LLM bot
that’s what I would say if I was an LLM bot!!
👀
I haven’t had any issues with my nvidia GPU. I did some distro-hopping and didn’t have any nvidia issues in any of the distros I tried.
If you want everything to work out of the box, I would recommend Bazzite. Pop! OS had me using the AMD image and fetching the nvidia driver manually (the nvidia image just didn’t work for me). After that, everything worked brilliantly.
Bazzite just works unless you have a Gigabyte b550 motherboard. Guess what I have?
the nvidia desktop version of bazzite didnt work with vulkan for me. it was still attempting to use mesa drivers for it. this was after debian where what i was trying to do required bleeding edge drivers which obviously wasnt going to work. then i just said fuck it and went with ubuntu like i have my entire linux career. you can hate on me, but it honestly works good enough
No hate, fam. Penguin brothers stick together
No way ? this is what I have. What are the issues ?
The issues were random black outs when the system was idle. The system just shut off display output and you had to force shutdown. Only logs that were there pointed to a popular Bazzite sleep issue. Didn’t look like it was worth it trying to patch it (fresh install) so I just swapped over to CachyOS.
ok cheers good to know
I think it’s more of a
Open drivers vs proprietary drivers
Well, can’t say for everybody, but i have no trouble running nvidia gpu on Hyprland with nvidia-open drivers. Haven’t spotted any troubles with Plasma or MangoWC either, even though i haven’t used them for as long.
I’m pretty sure you need the proprietary ones but I could be mistaken
I’m super annoyed at Fedora workstation at this moment. My 240hz Samsung monitor can’t use HDMI to get to 240hz, regardless of the quality of the cable. I have dual monitors and one is already using the type c so one of my monitors have to be 120hz.
That’s the HTMI forum for you
Never had an issue with my Nvidia card. OBS can use the hardware encoder out of the box. Just a few weeks ago upgraded to a AMD card and had to set some “advanced” settings in OBS to do the same. Really happy overall, but after seeing this meme for years I expected rainbows and sunshine but was unpleasantly surprised in that regard.
my nvidia card caused sleeping and hibernation to randomly and regularly fail, and it made me very vary of system updates breaking random things.
My nvidia card prevents suspend working properly, but to be fair my previous nvidia card had the same problem when it was in a Windows machine.
I never use sleeping or hibernation, so can’t attest to that functionality.
Updates never broke random things for me with regards to the gpu. My install is 7 years old, so it’s been updated a lot.
My nvidia card causes horrendous screen tearing in VR if my monitor supports variable refresh rate. I have to unplug the gaming monitor to use VR
I don’t have VRR monitors and only occasionally dabbled in VR, to my experience without issues besides ALVR disconnecting from SteamVR sometimes. I picked up the VR set now that my system is beefed up and I still have the same issue sometimes, so I’m not chalking this up to my older Nvidia card or drivers.
When you want to do GPU processing for AI, crypto, video editing, etc, though, this gets reversed.
Getting Cuda working on Linux with an nvidia card is relatively painless. Just a few well-documented commands, worked on the first try.
I could never get AMD’s equivalent to work on Linux, though, and it led me down a horrible rabbit-hole of trying a dozen different driver versions from a dozen different places, all with their own unique and quirky ways of installing… And it still never did work.
Thats just poor distro support, kind of like CUDA in the past… ROCM should “just work” if it’s shipped right. But it’s not really a priority with maintainers.
Now, if you’re trying to run CUDA stuff with ROCM, that’s a whole different story. The bast majority of GPU software has extremely poor ROCM support compared to CUDA, and some of this is definitely from AMD footgunning.
For me it was deadsimple once i tried setting it up with nix, granted you need to learn a little about nix so maybe that cancels it out a bit lol.
For me cuda was painful. I did the well documented commands, rebooted and had no output on my laptop screen anymore. Probably a complication due to Optimus, but still…
It was definitely Optimus. If you’ve got an Optimus laptop, everything bad in your life can somehow be traced back to it. Bad battery life? Optimus. Buggy video? Optimus. Hurts when you pee? Optimus. God I fucking hate Optimus.
ROCM still barely works on Windows and it’s only recently been supported at all IIRC.
ROCM is pretty simple. It’s just no where near as robust and supported as Cuda.
I run a legacy NVIDIA graphics in my ten year old laptop. GeForce 750M. The proprietary drivers are faster and have real video acceleration but haven’t been updated in forever and don’t support Wayland.
Nouveau works okay. I haven’t gotten video acceleration to work yet, even with installed firmware. Nouveau-vulkan is a bit buggy.
I’m running wayland with nvidia-open and nvidia-utils packages, and have never encountered any driver issues in both graphics and compute.
The only time I’ve ever really had issues with Nvidia drivers is when installing the meta package for CUDA (because it tends to include a previous version of the driver, which causes install/uninstall havok), or with laptops and hybrid graphics.
But the laptop issue is almost completely gone with newer distros like Bazzite.
because it tends to include a previous version of the driver, which causes install/uninstall havok
To be fair, this is a packaging/distro problem, as CUDA should always work (and be kept in sync with) the newest graphics driver.
ROCM and OpenVINO (AMD and Intel) are even more of a pain, actually.
You mean all three apps that support waylamd are working? Wow. At the same time?
Keep living in the past old man.
