The Linux Thread - The Autist's OS of Choice

  • Want to keep track of this thread?
    Accounts can bookmark posts, watch threads for updates, and jump back to where you stopped reading.
    Create account
I am so fucking tired of hearing the same disingenuous argument about how "none of my games work on Linux" and, when you finally get them to explain which games, 99% of the time it's live service competitive shooter slop where the wants to install some glorified fucking malware.
 
Day of the Cope said:
I am so fucking tired of hearing the same disingenuous argument about how "none of my games work on Linux" and, when you finally get them to explain which games, 99% of the time it's live service competitive shooter slop where the wants to install some glorified fucking malware.
Click to expand...
I just wish these people would be honest, and say that they don't want to learn something new or be inconvenienced in any way with Linux. And just to stick with their familiar Windows experience.
Outside of these multiplayer games with horrible anti cheats, every other game in my Steam library runs just fine under Linux. But you'll never hear people talk about the 99% of games that work, only the 1% that they had trouble with.
Linux is a fun time but its fundamentally different to Windows. Which I see a lot of people struggle to grasp, but only with Linux.
Normal people are able to switch from Windows to Mac and vice versa just fine despite being totally different.
 
Day of the Cope said:
I am so fucking tired of hearing the same disingenuous argument about how "none of my games work on Linux" and, when you finally get them to explain which games, 99% of the time it's live service competitive shooter slop where the wants to install some glorified fucking malware.
Click to expand...
Gaymers are fat on cheetos and are too poor for Steamdeck, so if they sit on their blubber gut all day complaining, then that's just a normal day, being fat virgins with rage.
XLibre should have a dedicated distro.
 
Akerman said:
Found this ancient thing from '93 after snooping about. Apparently it inspired dpkg:
View attachment 7808115

StopAlop was a Bourne-style shell script that builds, installs, lists, verifies, and removes binary packages using only core Unix tools. It is largely a relic of early Linux days when tools like floppies and manual file tracking were common. As for the meaning behind the name:
View attachment 7808138
:story:
Click to expand...
Early description of "malding", exemplified by many characters in the 40K universe
40k-evolution-of-horus-lupercal.webp
Slav Power said:
Must've been an avid reader.
View attachment 7808196
Seriously, how come there is not a single high resolution version of this cover available online?
Click to expand...
Found this on Twitter, though it will require a little editing:
unix-haters-handbook.webp
I fuckin' love that book. Know what else I'm a big fan of?
bofh.webp
The Bastard lives on even today at The Register:
 
Day of the Cope said:
I am so fucking tired of hearing the same disingenuous argument about how "none of my games work on Linux" and, when you finally get them to explain which games, 99% of the time it's live service competitive shooter slop where the wants to install some glorified fucking malware.
Click to expand...
Changing to Linux has helped me avoid having to buy the latest AAA slop just because someone in the friend group had a crave for it that lasts a week or two.
 
Premium Dog Food said:
Only game that runs better on linux than windows is minecraft, before adding any optimization mods.
As personal experience, games run on about the same performance on linux than on windows, except on linux I don't got to worry about windows update shitting my computer while gaming.
There isn't any real gain in performance other than not having windows services doing whatever the fuck in the background.
Click to expand...

Factorio is better on Linux, better performance and more importantly non-blocking saving (You can keep playing while the game saves, on windows it pauses for a bit)
 
Esope said:
i think it's more like "none of the cracked games works"
Click to expand...
funny. You would think with how the game industry uses "cheating" as a reason they need to block linux support. You would think cracked games would be easier to use on linux. Not exactly the same as cheating, But still.
 
Betonhaus said:
I genuinely thought that as Xlibre became available on more distros interest in it would snowball faster. I'm sure it'll still build interest over time.
Click to expand...
It'll look more like non systemd init adoption and serve as one half of an ideological dipole. X versus Wayland has produced some very ugly opinions.
 
Esope said:
i think it's more like "none of the cracked games works"
Click to expand...
I've had very bad luck with Fitgirl repacks, DODI works most of the time, and GOG's installer can be kinda buggy but will work. Nowadays I just use Goldberg emu and clean steam files from cs.rin.ru. It is the most reliable method, it's where all the repackers source their files from, and you don't have to deal with compression bullcrap.

I can't remember the last time I've had a "cracked" game I couldn't play.
 
Betonhaus said:
I genuinely thought that as Xlibre became available on more distros interest in it would snowball faster. I'm sure it'll still build interest over time.
Click to expand...
This is the exact reason. Much earlier in this thread, I said it was so bad that it started, out of the gate bringing up DEI in the readme by name. He could have even said the thing right after, about everyone being welcome without saying dei, it would have meant the same thing. And it would have stopped the people that immediately went against the project for purely political reasons.

Yes they would have still found other things in the guys past they didn't like. But when it's not directly tied to the project itself. It's easier for people to "separate the art from the artist"

That and lunduke. Immediately made his video saying it's an anti-dei project. As if that's the only point in it existing. Basically saying this is a political project. Rather than a technical one. That's what most people especially ones that aren't paying attention to the identity politics shit. Or are left leaning, think.

And when most people see anti dei, think that means they are saying something completely differently than what they meant by it.

I talk with people outside of this thread. And see what they think of the project. And they are reacting exactly how I thought they would react. A lot don't want anything to do with it, or have a negative perception of the project. Purely because of politics. And it absolutely is going to make this projects life much harder, when it was already going to be an uphill battle, just from the goal of the project from the onset.

Its unfortunate because I do like that it exists, and I do want to see it succeed. And the "it keeps out the people you don't want contributing" thing. Really doesn't make sense here. Because for this to actually succeed, you need as many people as possible using it. Otherwise it's just not going to be supported. xorg, isn't like some small random project, where, if trannies don't want to use it, it has no consequences. No one wanting to work with your project in the case of this. Means a slow death to the xorg ecosystem. It means distros won't want to package it, and make your life harder. It means less eyes on it to catch bugs. If you care about xorg continuing you should want as many people using this as possible.
 
I finally had a chance to write up a simple test regarding NVIDIA's drivers on Linux vs. Windows that anyone can mess around with.

Steps to test
  1. Open a bunch of applications you'd normally use throughout the day (for later tests)
  2. Pick any three decently heavy video games (e.g. GORI, Death Stranding and Cyberpunk 2077)
  3. Run them and load into any level you wish, so that each game fills dedicated VRAM decently
  4. Suspend/pause the first two games, so that they're not consuming any GPU processing time
  5. Observe the subjective performance of your singular running game with the other two suspended
  6. Try switching between games by suspending/resuming their processes, observing the impact
  7. Observe the effect under load when switching between existing running applications
  8. Try opening other apps while the system is in this state to make sure everything works as usual
Relevant Commands

PurposeLinuxWindows
Monitoring Resource Use
(for NVIDIA cards)		Use the CLI to get reliable stats:
Bash: 
watch -n 100 'nvidia-smi -q |
grep -i "FB Mem" -A 3 && nvidia-smi -q |
grep -i BAR -A 3 && nvidia-smi |
grep -E "[0-9]{2,3}MiB \|" |
grep N/A && free -m'
Just use Task Manager on Windows 11

nvidia-smi can't fetch usage stats because of WDDM and unfortunately Task Manager values aren't perfect because of the abstraction layers the GPU drivers are forced to utilise.
Suspending Processes
Code: 
pkill -STOP <name of process>
Code: 
pssuspend <name of process>
Resuming Processes
Code: 
pkill -CONT <name of process>
Code: 
pssuspend -r <name of process>
Terminating Processes
Code: 
pkill -9 <name of process>
Code: 
taskkill /f /im <name of process>

Linux Results (NVIDIA)

VRAM usage:
Code: 
FB Memory Usage
        Total                             : 12282 MiB
        Reserved                          : 163 MiB
        Used                              : 11991 MiB
BAR1 Memory Usage
        Total                             : 16384 MiB
        Used                              : 4654 MiB
        Free                              : 11730 MiB

The top consumers of VRAM were:
Code: 
|    0   N/A  N/A            2785      G   /usr/bin/kwin_wayland                   306MiB |
|    0   N/A  N/A            2945      G   /usr/bin/plasmashell                    501MiB |
|    0   N/A  N/A            3899      G   /usr/lib64/firefox/firefox              315MiB |
|    0   N/A  N/A           19884      G   ./steamwebhelper                         60MiB |
|    0   N/A  N/A           42230    C+G   ...dlyCarnage-Win64-Shipping.exe       3889MiB |
|    0   N/A  N/A           44337    C+G   ...TRANDING DIRECTORS CUT\ds.exe       4336MiB |
|    0   N/A  N/A           46589    C+G   ...077\bin\x64\Cyberpunk2077.exe       1620MiB |
|    0   N/A  N/A           49283      G   /usr/bin/systemsettings                 111MiB |
|    0   N/A  N/A           56675      G   /usr/lib64/firefox/firefox               63MiB |

The NVIDIA Linux drivers are completely incapable of evicting anything from dedicated VRAM in this scenario, even when processes are suspended. The third game I launched (Cyberpunk 2077) ended up with insufficient VRAM from the GPU and ran like a slideshow.

Existing Wayland-native processes would mostly remain working (except for launching a second Firefox profile). New Wayland-native processes would either crash out, render transparent windows or somehow find a way to snake their way into what little framebuffer memory remained. XWayland processes would always end up rendering transparent windows when launched, even things with software rendering like the GTK2-powered Nero Linux. Upon closing one of the games, applications would begin to work, but some needed a manual resize before their contents would redraw.

One thing I haven't tried is sticking with purely open-source native Linux games designed to work natively on Wayland to see if the end result is the same. But that's mostly because I can't think of any which are heavy enough to challenge a 4070 Ti.

Windows Testing (NVIDIA)

VRAM usage:
1755811189648.webp


The top consumers of VRAM were:
1755812093979.webp


Cyberpunk 2077 was as playable as it would normally be (only the odd occasional hitch or two, usually along with pop-in when streaming in assets from other areas). Suspending/unsuspending processes to switch between games yielded similar general framerates to when they run standalone. The drivers also automatically reclaimed dedicated VRAM from any processes not visibly on-screen quite aggressively. The most notable issue under this level of load was a small amount of latency switching to/from applications which had been completely shunted out of dedicated VRAM, where the system would stutter for a split second (including the mouse pointer) but to be fair, this also happens from time to time when querying monitor/display info or loading some of the Task Scheduler UIs under normal load anyway.

Just to see what would happen under absolutely ridiculous VRAM pressure, I did try loading in a 12b model using koboldcpp with all 40 layers GPU accelerated, in addition to the above. As the rest of the games plus other software had already consumed a large amount of RAM, as well as both dedicated and shared VRAM, the NVIDIA drivers no longer had easy access to further system RAM to repurpose. This resulted in a very stuttery desktop, with no improvement after 3 minutes of waiting. Suspending the koboldcpp process also did nothing to improve this, as it would jump between 0 and 100% GPU usage irrespective of whether it was actively running or not, suggesting the NVIDIA driver was likely backlogged trying to scavenge for resources. Ending the koboldcpp process immediately resolved the problem, however, and even the games started working again.

A silly thing I haven't tried is using RAMMap to manually empty out working sets, standby pages etc. to free up further resources for the driver to claim, but that will have to wait for another day.
 
prollyanotherlurker said:
funny. You would think with how the game industry uses "cheating" as a reason they need to block linux support. You would think cracked games would be easier to use on linux. Not exactly the same as cheating, But still.
Click to expand...
My theory is that any cracks that rely on DLL swaps, like Goldberg's Steam Emu, or some other DLL linking, will need to be worked on manually as under Windows it's just a drag and drop, but under Wine you have to explicitly declare DLL swaps. It's also an issue with modding where using an ASI loader is common, but what it is is a DLL file that utilizes common "open entry points" in processes, such as dinput8.dll, so that other DLL files renamed to ASI can piggyback off of it and modify the game at runtime. Again, on Windows it's all drag and drop by design, but on Linux it's explicitly declared by design.

Otherwise a cracked game is a cracked game, some of the game files got slightly altered to bypass the DRM checks but Wine shouldn't really care about it since it'd treat those modified files the same way it would treat the original ones.
Perfectly Innocent said:
A silly thing I haven't tried is using RAMMap to manually empty out working sets, standby pages etc. to free up further resources for the driver to claim, but that will have to wait for another day.
Click to expand...
There's also this piece of software called MemReduct that does those flushes automatically.
 
Esope said:
i think it's more like "none of the cracked games works"
Click to expand...
Yandex Captcha Solver said:
I can't remember the last time I've had a "cracked" game I couldn't play.
Click to expand...
Same here, my 6 year old Ryzen/Radeon rig could pretty smoothly handle Cyberpunk 2077 and Baldur's Gate 3, both from Dodi and both with a whole bunch of mods, easily through Lutris/Nobara. Not sure what happened but trying to install a Fitgirl repack of the latter went so poorly I had to reinstall the whole OS by the end of it. Could not get past TTY before it froze and died. Whatever black magic they're doing apparently doesn't play well with Linux.
 
Perfectly Innocent said:
I finally had a chance to write up a simple test regarding NVIDIA's drivers on Linux vs. Windows that anyone can mess around with.
Click to expand...
You can try running Proton under Wayland as well to circumvent Xwayland and see if that makes games run better under such a scenario (even if it would still be a problem) with PROTON_ENABLE_WAYLAND=1 %command% in launch options with a version of Proton that supports it. I suspect it would then give you the same behavior as apps with native Wayland support. wine-wayland is still sorta experimental but I don't notice as many issues running games under it though some games do play worse. There is no Steam overlay but that's the only regression I have noticed consistently. I don't care about Steam overlay personally.

In fact it actually fixes some games that don't play well on Xwayland because Xwayland is buggy for me in so many ways on NVIDIA, for example I got a memory leak on Firefox that ended with it OOM killing itself after filling my swap. That didn't happen on Firefox under Wayland.
 
Last edited:
Dread First said:
Honestly, not even gonna bother putting Fedora 42 KDE through its paces on Wayland; if Deus Ex GOTY and Dark Souls PTDE shat themselves on Wayland under GNOME shell, I damn sure ain't gonna bother trying again on Fedora 42 KDE.
Click to expand...
If gaming on Fedora 42 KDE is a priority, why not try running Nobara? It is made by the guy behind GE Proton so I would expect it to have better game compatibility.

Slav Power said:
This is why I have a problem with all of these Windows vs Linux benchmarks. They're not objective, they're not informational. They'll cherry pick a few select games and twist the results in such a way that they can claim Total Linux Victory and it's so disingenuous that it's just adding fuel to the dumpster fire that is "Linux community's reputation in the common consensus".

What I'd like to see, is a massive test of all kinds of games. New, old, DirectX 9 to DirectX 12. Vulkan and OpenGL. All of them done on a decent PC rig, something with a fully sized RTX 3060 and an AMD equivalent to compare how both GPU vendors are handling themselves under both OS'. A massive dataset where you can't do some smarmy journalistic data manipulation, just an objective megachart that'll show how every game compares.
Click to expand...
The comparisons done by Phazer Tech on YouTube seem to be pretty objective. The sample size of games isn't great, but I like his approach and I think he is being fair with his conclusions.

AMD video card comparison video:
On average, Windows tends to have framerates 7% higher than Linux at 1080p, and 5% faster at 1440p.

NVIDIA video card comparison video:
On average, Linux is 14% slower than Windows at 1080p, and 16% slower at 1440p.
 
You must log in or register to reply here.
Back
Top Bottom