As someone who hasn’t tried to game on Linux but just use it as a server with my old pc hardware, dealing with nvidia shit is just a massive pain in the ass.
I was only using one monitor and yet it’d never pick up edid properly and other random quirks.
I chucked an Intel arc in there for av1 encoding on jellyfin and after getting to kernel 6.2 it also “just worked”.
It’s amazing how much of a difference it can make when the manufacturer gives even one quarter of a shit about Linux.
I started dual booting Linux back when Steam for Linux was reasonably new and Portal 2’s native port was on beta. Briefly went back to Windows after building a new, much powerful system for about a year, DXVK & later on Proton happened, and now all the games I care about work flawlessly.
There have been games on my Steam library that I never ran on Windows despite them not officially supporting Linux.
With the deck I seriously hope devs slowly but surely start thinking about native ports as well, but I won’t mind waiting another - uhh, 10?! - years for that to happen. I expect Steam Linux Runtime & Flatpak to be the DXVK & Proton of native ports - as in, the thing that will make them “viable” instead of “theoretically possible”. Win32 is still the most stable ABI on Linux after all.
Ahh thanks this looks promising. The only thing that I don’t get from that article is where they talk about the nvidia gui settings. I’m not seeing that. Is that the same as the nvidia xserver application?
So I found the issue in case anyone comes across this. I had to switch to x11 gnome instead of wayland on the login screen in fedora 38. After that Diablo runs like butter!
i got a steam deck a few months ago and am constantly amazed at how well it performs. in fact, assassin's creed 2 plays better on the deck than it does on my seven-year-old gaming rig
needless to say, once windows pulls the plug on 10 i'm fully converting to linux and not looking back
10 years, that’s a long time ago! It’s mostly in the last 3-5 years that things started getting really good with Vulkan becoming a thing and DXVK being made. DXVK is really impressive how fast it got put together and how drastic the improvement is over wined3d.
This video (series) is so cathartic. Love when people reach that moment of frustration with a company and break it off completely, instead of just eternally bending.
Get yourself a splitter for the 4 pin fan cables. Disconnect one set of fans from the controller. Connect the unplugged set of fans directly to the splitter. Connect the controller to the splitter. Connect the splitter to the motherboard.
Your bios should now see your fans. You can create custom fans curves and dynamically control the speed of the fans.
It may not solve the problems you’re seeing, but it seemed worth sharing.
Thank you so much for the advice but the issue with the Lian Li Uni Fan v2 is that the fan connectors are proprietary and only fit into the controller. The only thing I could split is from the controller and I don’t think that would really change my current situation.
Sorry for the delayed response - I saw what you meant in the v2 manual about the proprietary connector and thought I would just leave it at that, but I had another thought.
It sounds like you do have some access to Windows/L-Connect, have you switched the fan profiles to sync with the MB? This should stick after a shutdown and not need to be re-set, you will lose the ability to create fan curves in L-Connect: https://lemmy.ml/pictrs/image/fac93968-ab47-407b-8e49-91577d8762d9.png
Also, are you connecting the UNI HUB directly to the MB pins, or is it running through a splitter or extender? It looks like the UNI HUB v2 does have a sense pin (the v1 does NOT, that’s the reason for the workaround I have to use); some additional extenders and splitters are missing the sense pin, so even though you can control the fan speed from BIOS, you can’t see the RPMs in BIOS so it might look like it’s not recognized.
Thank you for the continued effort to help! If I boot into windows and enable the “MB RPM Sync”, the fans are recognized by the BIOS and ramp up and down as needed. This setting does not persist after a shutdown tho. The only way I’ve found to consistently have the fans work as intended after a shutdown is to unplug the USB cable from the Lian Li controller, which kills RGB. I have the PWM connector from the controller going straight into one of the connectors on my mobo. I can tell that after a shutdown, the fans just rotate on their lowest setting, until I open L-Connect 3 again on windows or disconnect the usb port. I’m attempting to implement this: github.com/kryzaach/Lian-Li-Unifan-Sync but think I have it placed in the wrong location, as it errors out immediately.
Geeze, that’s extremely frustrating… they really switched up the headaches between v1 and v2. I’m sorry I couldn’t be more help! Hopefully you can find a software solution to get it going.
I don’t mean to repeat troubleshooting you’ve already done with the python script, apologies in advance if this is stuff you’ve already done. If it’s erroring immediately, you may want to double check that you have pyUSB in place and UDEV rules to get access to the device if you’re not running the script as root.
edit: You may also need to check that your Vendor ID and Product ID match what the script was written for, given that there seem to be some different iterations of Lian Li’s hardware just in general.
Thank you so much for continuing to try and help. It’s greatly appreciated. I am very new to Linux so while I can do some basics, this is very helpful. I installed pyUSB but still get an error when running the script. I have the script placed in my home folder. Should it be elsewhere? The error I get is “No module named ‘usb’. I did adjust the script to have my controller’s proper identification so feels like I’m missing a small piece of the puzzle!
I’m not familiar with Nobara, but I can at least show how I would install pyUSB for use in Debian:
If you follow the instructions exactly as in the readme, it will error:
<span style="color:#323232;">$ pip install pyusb
</span><span style="color:#323232;">error: externally-managed-environment
</span><span style="color:#323232;">
</span><span style="color:#323232;">× This environment is externally managed
</span><span style="color:#323232;">╰─</span><span style="font-weight:bold;color:#a71d5d;">></span><span style="color:#323232;"> To install Python packages system-wide, try apt install
</span><span style="color:#323232;"> python3-xyz, where xyz is the package you are trying to
</span><span style="color:#323232;"> install.
</span><span style="color:#323232;">
</span><span style="color:#323232;"> If you wish to install a non-Debian-packaged Python package,
</span><span style="color:#323232;"> create a virtual environment using python3 -m venv path/to/venv.
</span><span style="color:#323232;"> Then use path/to/venv/bin/python and path/to/venv/bin/pip. Make
</span><span style="color:#323232;"> sure you have python3-full installed.
</span><span style="color:#323232;">
</span><span style="color:#323232;"> If you wish to install a non-Debian packaged Python application,
</span><span style="color:#323232;"> it may be easiest to use pipx install xyz, which will manage a
</span><span style="color:#323232;"> virtual environment for you. Make sure you have pipx installed.
</span><span style="color:#323232;">
</span><span style="color:#323232;"> See /usr/share/doc/python3.11/README.venv for more information.
</span><span style="color:#323232;">
</span><span style="color:#323232;">note: If you believe this is a mistake, please contact your Python installation or OS distribution provider. You can override this, at the risk of breaking your Python installation or OS, by passing --break-system-packages.
</span><span style="color:#323232;">hint: See PEP 668 for the detailed specification.
</span>
Instead, I would create a virtual environment (I’m using virtualenv instead of venv)
But why does it end up washing out colors unless I amplify them in kwin? Is just the brightness absolute in nits, but not the color?
The desktop runs in SDR and the color space differs between SDR and HDR, meaning you will end up with washed out colors when you display SDR on HDR as is.
When you increase the slider in KDE, you change the tone mapping but no tone mapping is perfect so you might want to leave it at the default 0% and use the HDR mode only for HDR content. In KDE for example, colors are blown out when you put the color intensity to 100%.
Why does my screen block the brightness control in HDR mode but not contrast? And why does the contrast increase the brightness of highlights, instead of just split midtones towards brighter and darker shades?
In SDR, your display is not sent an absolute value. Meaning you can pick what 100% is, which is your usual brightness slider.
In HDR, your display is sent absolute values. If the content you’re displaying requests a pixel with 1000 nits your display should display exactly 1000 nits if it can.
Not sure about the contrast slider, I never really use it.
Why is truehdr400 supposed to be better in dark rooms than peak1000 mode?
Because 1000 nits is absurdly bright, almost painful to watch in the dark. I still usually use the 1000 mode and turn on a light in the room to compensate.
Why is my average emission capped at 270nits, that seems ridiculously low even for normal SDR screens as comparison.
Display technology limitations. OLED screens can only display the full brightness over a certain area (e.g. 10% for 400 nits and 1% for 1000 nits) before having to dim the screen. That makes the HDR mode mostly unuseable for desktop usage since your screen will dim/brighten when moving large white or black areas around the screen.
OLED screens simply can’t deliver the brightness of other display technologies but their benefits easily make it worth it.
Ah cool, I didn’t know that there are layers of capabilities for different requested brightnesses. Thanks for your in depth reply! I’m also a 1000 nits enjoyer but I don’t switch on any lights - I like when my eyeballs get blasted with colors. 😂
I have done things to this game that morals prohibit.
Like copying a whole Skyrim main folder modded with Nexus Mod Manager from a laptop to a steam deck and “managing” this with MO2. Bonus obscenity: couldn’t run the new Skyrim version instalmed on Steam Deck with the old SKSE from the laptop install, so I had to update SKSE, break compatibility with the load order, and then manualy find which of the mods to update for it to run. Do note that I have 250 mods and an historic 70+ level build on this instance.
I guess what I am saying is “you can go pretty wild with this”. Start anywhere, end up anywhere, mod the shit out if it.
I would advise on following one of the “FULL OVERHAUL NOTHING UNTOUCHED” modding guide, just because why not. Most of them have some form of attention to performance that will help with your issue.
I used mo2 from this installer github.com/…/modorganizer2-linux-installer.
Both on Nvidia and AMD it’s been great. Older version had a bit laggy mo2, but now that is gone. Only issue I have is that ModOrganizer itself needs to be installed to main drive, otherwise it wouldn’t work, but maybe it’s also fixed now.
I use it for Starfield at the moment and it works okay. There's definitely some caveats though and it's still something that might require some tinkering / fixing. That said, the last time I tried Skyrim on Linux it also ran atrociously bad, the framerate was just not what I'd expect from my system and way worse than what I had on Windows with my previous system (which was much worse in). Similar experience when I tried Fallout NV. The performance dipped down into the low two digits and I didn't even really went into modding all sorts of stuff into either of them.
Starfield runs much better but I suspect it might still perform better in Windows. Not sure what it is but it seems that gamebryo just does not run well under Linux, or rather even worse than under Windows.
Maybe I try NV again at some point since I kinda want to play it before season 2 of the TV show. It's been a while since I tried so maybe there had been some fixes & improvements since then and I've seen some engine optimization mod too.
I'd still hope for a native mod manager for Linux though. R2Modman is kinda decent but does not support Bethesda games.
There is a change for a cross OS malware but it’s probably still quite small.
I would just remove them, and then if they dont re-appear after checking game file consistency then its’s mostly likely not a false positive.
But if you know that last playerd is most likely before you installed current OS then it should be clean. I would check user level autostart and cronjob (so mostly stuff in $HOME/.config) places if something didnt get added, go ahead with life.
Unless you are doing banking and other critical stuff on this machine then I would be thinking of hardedning the OS in some way and/or reinstalling.
Given that Priton / Wine / ... try to simulate a Windows environment, it's certainly possible. But I've never heard this happened to anyone. And the discussions on the internet seem very theoretical.
Surely it highly depends on the exact kind of malware and what it tries to do. And it doesn't really seem to happen in the wild. But I see no reason why it wouldn't theoretically work.
linux_gaming
Top
This magazine is from a federated server and may be incomplete. Browse more on the original instance.