For whatever reason, Windows 11 is worse at Cyberpunk 2077 than Arch for me. Constant stuttering. It might be that Arch has much less going on than Windows, but it’s enough for me to use Linux as my main gaming OS now.
Real question- I have a steam deck and am incredibly pleased with the playability. I also have a desktop with a newer nvidia card. Does Linux have support for DLSS yet? It make a huge difference in oerformance and honestly it’s the only thing holding me back
It should support DLSS unless you have an older video card, which the drivers don’t work well with. I heard the newer Nvidia cards work better though. Of course, is all up to you whether you like it or not, so just try out Linux and see. If you don’t like it just reinstall Windows. Make a recovery Windows USB beforehand though, makes it easier to reinstall.
That depends which DLSS. In my testing DLSS 1 and 2 work fine in games that I tried, with recent Proton enabling it as well as ray tracing shouldnt require extra steps anymore (it was experimental and opt-in using environment variables). DLSS 3 with frame generation is known as no go yet and it’s unfortunately on NVIDIA to provide support for it as it’s very much locked down guarded proprietary stuff.
Linux and Nvidia don’t mix well, at least not until Nvidia’s official open source kernel module has been upstreamed to the Linux kernel which will take years.
Breakages, workarounds for breakages, etc. are common occurrences, especially when you want to use a modern desktop using Wayland.
Other than being completely unable to run Wayland, secure boot, and being forced to use a propietary driver what kind of things are specifically wrong with Nvidia on Linux? Maybe it’s because I switched to Linux fairly recently but I haven’t noticed many Nvidia specific issues yet.
If you want to game, stick to regular Fedora. A project that is actually secure is ublue with dedicated NVIDIA images that should just work and never break, and they even have Bazzite, an Image specifically for the Steamdeck but also for Desktop.
These images are only ½ day behind upstream, apply minimal additions and patches (like drivers, codecs, packages, udev rules for controllers) and Nick from the video above found out that the Nobara patches with their weird less supported Kernel arent really worth the hassle.
Proprietary UEFI BIOS is, but for a secure system with local manipulation prevention it can be needed. Also secureboot is a security measurement against malware so no, its simply the best we have.
Look at Coreboot if you want a secure modern system
The vulnerability actually isn’t in Windows Boot Manager, it’s a flaw in the image-parsing code of the UEFI itself. That’s why it’s able to bypass SecureBoot.
It just happens that for whatever reason you can easily update the image file from within Windows/Linux itself. The fact they don’t show a logo currently does not mean you’re immune, as the system might just be showing a black screen at that point. Code can be injected into an image file without perceptibly affecting the image output, so you’d likely be able to use a “black screen” logo. If your computer has a UEFI instead of a BIOS, which is pretty much everything from the last 10yrs, then you are more than likely at risk.
My computer likely isn’t susceptible, and that’s because it’s a Dell workstation. While the bug still exists in the image parser, Dell has locked things down so it’s pretty much impossible to change the boot logo from userspace.
I 100% agree, its best to just stick to upstream Fedora imo. Glad you made this comment. The security issues of Nobara always put me off, especially since basically everything it does can just be applied to regular Fedora. I think Nobara would much better serve as a script or toolkit, similar to Brace, or something along those lines instead of an entire separate OS with the security issues it brings.
And that definition depends on how you define “benefitting the user”. If someone has an online match ruined by a hacker, I’d argue that they would have benefitted from the game running some kind of anticheat.
Do we define user as the singular individual person? Or do we consider the user as a collective, and factor in the larger benefit to the masses? It could even be argued that the people running cheats are the ones running malware (specifically, malware that targets the other users in the match) and should therefore be treated the same way we treat people who use more traditional viruses and trojans at the detriment to others. The same way you wouldn’t want some virus-ridden machine connecting to your home network, (you’d probably want everyone to at least be running a basic virus scanner and have common sense when browsing,) you would want everyone in the game running anticheat to ensure there is no malware.
Very few people would say that it’s okay to waste others’ time and computer resources on a bitcoin miner trojan… Most people would (correctly) determine that it is theft. But then when it comes to online games, the same people feel entitled to waste other peoples’ time and computer resources by ruining their matches.
That’s largely a corporate decision that is out of the hands of the programmers. Generally speaking, security specialists would agree with you. But running anticheat on the server costs server resources, which means you need more servers to accommodate the same number of players. Running it client-side is a cost cutting measure mandated by the corporate bean counters who did the math and concluded it’d be cheaper for the company to spend the users’ computer resources instead.
While I agree that client-side security isn’t the best solution, it’s certainly better than no solution. It’s the same argument people have against self-driving cars. The self-driving cars don’t need to be perfect; They just need to be better than the average driver. If they can reduce the number and severity of accidents that are currently happening without them, then they should be implemented. Even if the solution isn’t perfect. Because an imperfect solution is better than doing nothing at all.
You’re right and it’s a pragmatic approach to the problem. They only need broad technical effectiveness to change user behaviour.
I’d argue that it’s not strictly cost cutting but cost transferring. The total client resources most likely exceed that which would be needed on servers.
I don’t think that is a widely accepted holistic definition of malware. But even if, AC is not waisting resources. It’s taking the resources it needs to perform its job.
You’ve linked to their anti cheat which they also offer but it’s not their main product. Funny that you missed that, given that you were already on their web site and irdeto.com/denuvo/ spells out “Anti-Piracy technology” in huge font:
There are games with single player and multiplayer modes that come with anti cheat. I had some game a few months ago that was a Steam freebie (can’t remember the name) whose anti cheat didn’t install properly on Windows and it didn’t allow me to launch regular single player, only mod mode.
There are several forms of anticheat. The ones that just run when the game is running, is usually fine. However, there is the Riot anti cheat which just runs all the time and isn’t uninstalled when Valorant is uninstalled. That is malware.
According to them ~58% of anti-cheat games work. There’s been a large uptick of anti-cheat support since the Steam Deck.
According to ProtonDB, 86% of the top 1000 games on Steam function (Silver+ rating). It’s a pretty safe bet that the most of the missing 14% is probably due to anti-cheat.
Wait, isn’t a lower frame time better? Why does their screenshot show windows having the lowest and say that it scored last?
Looking at the source article, windows did have generally better 1% lows except for Starfield, so I think this article has it backwards. They also cherry picked 2 results where windows was worse lol.
I’m all for pro-linux stuff but articles like this just reek of making shit up so it looks better.
They probably didn’t label their axes properly. FPS is a clearly defined metric, and there, more is better. This indicates that the conclusion (Linux is faster) holds. Since frame times have an entry with value “100” and all other values are lower, I assume that’s in percent, i.e. Arch Linux is the fastest and picked as comparison point, and the others are shown with relative performance to Arch.
I think FPS was actually selected, not frametimes. 1% low frametimes of 89 does not make sense.
There is an issue with the image in the article, but not the one that you might think it was. The FPS should have been more clearly indicated that it was the selected tab and then it probably would have been fine.
edit: I went to the base website www.computerbase.de/2023-12/…/2/ it’s in German, but, it seems like the frametimes and frame rates are nearly the exact same values - which doesn’t even seem to make sense to me?
A typical Linux distro, especially lightweight and simpler ones like Arch, will of course be better than a bloated OS, like Pop or Windows. The only problem with Linux distros might be the choice of tools - X and AMD will work much better overall than Wayland and Nvidia.
Just that many people may have an Nvidia GPU before deciding to use Linux, and some people just prefer to use Wayland over X for literally everything else.
My PC with Wayland + Nvidia has so many problems with gaming, especially flickering and performance, while my Laptop with Wayland + integrated Intel graphics has no problems at all - even in games, that I wonder if Nvidia + Wayland still really sucks ass or if my GPU is just broken. Currently there’s a bug where frames are ‘switched’ somehow, so it’s not Frame 1, Frame 2, … Frame n, but Frame 1, Frame 3, Frame 2, Frame 5, Frame 6, Frame 4 etc.
I expect it to be fixed by an update of nvidia in the future, but there are always such bugs.
If my experience is any indicator, your GPU is fine :(. Any chance you’re using mixed display scalings? I’ve got an RTX 3050 eGPU for my Plasma/Wayland laptop, and for the most part it actually works fairly smoothly (albeit more slowly compared to windows), but if I try to run a game at a higher resolution than my monitor (used by Plasma for mixed scaling) I get constant flashing/frame shifting, but when I drop it down to the native 1080p it starts working again
As a side note, X and eGPUs do not play well together, but Wayland is literally plug and play after installing the drivers–I can even hot plug/unplug as long as nothing’s using the GPU!
That frame issue is because of the fact that Nvidia uses “explicit sync” and AMD/Intel use “implicit sync” - XWayland is built to only support implicit syncing for now (Nvidia is trying to get it changed), and since most games right now run under XWayland… Along with a ton of apps of course.
Until then, that issue won’t be resolved sadly. It’s what finally pushed me to get an AMD card since the issue has been open for over a year with a ton of back and forth.
I’ll need to give Linux gaming another chance at some point.
All I know is that people were saying games run great on Linux a couple of years ago as well, but when I actually tried it for myself the performance was unusable.
Maybe that was my fault for over complicating my setup, but even when I tried a basic setup it still felt very janky.
Not sure if anyone’s able to advise, but does RTX and variable refresh rate work on Linux?
All three major GPU manufacturers support ray tracing and variable refresh rate on Linux. When playing windows games, ray tracing has to be handled through VKD3D, which AFAIK supports most but not all DXR features. I haven’t had any problems with it though.
The one thing that can still completely make or break your (Windows games on Linux) gaming experience is anti-cheat software, since it’s up to the game developers to enable it for wine. The major anti cheat providers offer solutions for this, but not all game studios are interested in their games running on platforms other than windows. Games like valorant will probably never work. Good riddance though.
Thanks, I’ll definitely need to give Linux gaming another shot then.
The last bit that might hold me back is getting my Hue Sync stuff working. It sounds silly, but it really makes games feel so much more immersive that I don’t want to be without it.
There’s a GNOME extension called HUE lights that allows you to control everything from your tray, entertainment zones and all. Similar probably exists for KDE/etc.
What about hdr. I saw it mentioned for the Steam Deck update, so wondered if that is finally working on Linux. I do like taking advantage of HDR on the TV.
It’s in the early stages, but yeah you can do it in KDE Plasma if you’re prepared to jump through a couple of hoops (basically doing the same thing the Deck does)
Linux won’t have proper HDR support until mid-late next year.
That’s in the works still right now, steam deck has it and I think it’s possible to get it working on other distros but isn’t on by default in most I don’t think
Valorant is a fucking awful game with über ban techniques when you force quit a game for some reason, like needing to go to the bathroom in middle of game play.
I can’t understand anyone can accept such a thing.
I’m sure there’s lots of solutions, but Steam with Proton for any windows only games has generally worked great for me.
Where I encounter issues, the Lutris flatpak install has worked well for me.
Both I believe use wine, but it is probably easier use downstream solutions like the above when getting started, instead of learning wine. Not that there aren’t benefits to learning it, just in a immediate issues -> lets go back to windows VS it just kind of works pretty good comparison.
Steam having a fair number of games that are directly Linux compatible now days is nice too.
Same, I could not get a single game to run normally on Fedora Kinoite, AMD GPU, Wayland. Idk maybe amdgpu pro and x11? But xwayland should also work normally…
The distribution “managed by a single person” depends on hundreds of people working on different sofware to keep up. It’s not “one person doing better than the thousands of Microsoft employees combined” implication they are pushing
Windows 11 beat the linux distros by up to 20% in 1% lows which are argued as much more important by most tech reviewers. It wasn’t consistant at all which means that there was a giant margin of error.
I love linux and linux gaming has gotten radically better, but I am tired of tech “journalism” literally just cherrypicking, misleading, clickbait trash.
Not to mention the major hurdle for Linux gaming is anti cheat software being brought over. Too many games are 100% unplayable because the devs don’t allow their anticheat to be installed on Linux systems
It also doesn’t work. I know that’s what the parent comment said, but it’s a total scam at the company level too.
“Oh, server networking is hard to do right. Let’s do it client side”
“Oh, people are cheating. Let’s add anticheat”
Ensue 3 years of fixing network consistency bugs and playing whackamole with cheaters
I’ve developed games where the client is the source of truth, and games where it’s the server. It is almost always better to do anything that will be developed for more than a few weeks serverside.
Also from an engineering perspective it makes LOADS more sense as you can apply patches to the servers instantly vs. requiring the users patch the game themselves.
Oh, absolutely. I can’t believe we deployed web apps on IIS for instance. What a shitshow that was. If you can run the important bits on something predictable like linux with all the serverside tools that gives you, why wouldn’t you.
In the defence of client side AC; if the entire game runs on the server, then network delay makes FPS:es awful to play. Being able to trust clients and let them do hit detection is quite important in making online FPS:es responsive. In addition, cheats that remove walls/grass, highlight players or even autoaim are near impossible to detect server side. One could try to use heuristics and statistics but it would be difficult to tell the difference between cheaters and players who are just good at aiming and map awareness.
Honestly I can't say that I miss installing rootkits with terrifying privileges just to play games. I'd rather limit the privileges games have with Flatpak etc., not give them even more.
Yup. People always latch on to the “Sony (it was actually on Philips, who ran the disc factory that Sony had a stake in, but that’s just nitpicking) installed a rootkit on PCs in the 90s via CDs” and say about how awful that is, and they’re right, then they throw that out the Window and install more advanced rootkits filled with god knows what telemetry when they install games.
Sure but gaming is predominantly a social pastime. Meaning that most gamers will make the trade off between installing anticheat and not playing the game their friends are all playing, much like the overwhelming majority of people will trade privacy in favor of being able to send a message to friends on Facebook.
It doesn’t matter how much you value your privacy: most people don’t care and never will. So without the option to give away privacy to play the latest Ubisoft game they won’t be using Linux. Full stop.
I really wish valve would make this more clear on steam store pages. It says games are “unsupported” on steam deck due to anticheat when really it should say something like “The developer of this title does not allow players using the steam deck” so that people are more aware it’s not linux or valve’s fault
If you compromise your system with software that you don't know and potentially can introduce a backdoor (even involuntary via bugs), you have a rootkit installed.
If you don't trust it, don't install it with admin privileges. Maybe don't install it at all. Anticheat is a shady business. And mostly not owned by the company that produces the maybe trusted product to be protected.
“A rootkit is a collection of computer software, typically malicious, designed to enable access to a computer or an area of its software that is not otherwise allowed (for example, to an unauthorized user) and often masks its existence or the existence of other software.”
That’s the Wikipedia definition, in CompTIA Security+ the concept of the malware masking itself is quintessential to the definition of a rootkit. I hear this shit all the time from people on here who think anything that gets elevated privileges is a “rootkit” and hasn’t the slightest idea what the fuck they’re talking about.
“But you don’t know if it could install a backdoor!”
You don’t know if half the shit you install is doing that either, or is Easy Anticheat known for doing this in some official investigation? Did someone find out that Activision is deploying malware in ricochet?
If not, you’re operating on suspicion that you don’t harbor for other software without evidence, based purely on things you’ve probably just barely heard about.
You should notice that I use the word "trust". I install stuff on my servers and PCs from people who I trust. Why should I trust someone who makes an anticheat engine. Why should I have a reason to do that?
You should also understand that a kernel-level piece of code that can be updated is a very good rootkit. It contains all essential tools to modify hardware, kernel, install drivers, keyloggers etc. It satisfies the definition of "rootkit" very well.
One single piece of code is enough to be a rootkit.
Rootkits: The rootkit is considered to be a type of Trojan horse. Many Trojan horses exhibit the characteristics of a rootkit. The main difference is that rootkits actively conceal themselves in a system and also typically provide the hacker with administrator rights.
Yeah, the only time proton can actually outperform windows is when it spots a fundamental performance error that the app has made, and is able to optimize it out, AND no windows driver does the same. This is comparing Linux+proton at its best vs windows+native at its worst.
What we really want to see is Linux+native at its best vs windows+native at its best. Unfortunately, there aren’t a lot of demanding games that natively support Linux.
As pointed out, in Windows defence, it’s actually faster where it matters. And none of it is going to matter in adoption until every thing is supported 1-1.
The only reason we’re behind on adoption vs Windows as this point is that people who write software for Windows, don’t do it for GNU/Linux, or even publish specs in the case of drivers.
It’s not the OSes problem. It hasn’t been for a long time. It’s stubborn developers (mainly corporations like Broadcom, Nvidia and Epic). We shouldn’t need to write compatibility layers for completely foreign software to run, or write drivers to drive a megacorporation’s hardware, and those are both a monumental task, but the community continues to achieve it anyways.
A lot has been done and continues to be done by the community, and that’s great, but the real problem is the corporations who refuse to invest a little bit of their time in GNU/Linux support (and those who have an irrational vendetta against it).
Causes are a part of the reality. And when people go online and complain about how “lInUX SuXxx” because their proprietary Nvidia drivers didn’t work, and blame the OS instead of the company who is meant to be providing proper support for their devices or at least documentation for other developers to use, it plants the idea in people’s minds that the OS itself is simply inferior, which has connotations of it just being a bad system. Instead of “it will work perfectly when drivers are actually released by the manufacturer”. It tarnishes it’s reputation even after that particular device gains support, and that is another reason why adoption is low.
Hell, nVidia was actively working against having a working opensource driver reverse engineered by Nouveau. Linux is a thorn in their side and the only reason they somewhat support it today is that GPU compute works so much better on Linux.
Add comment