The summary includes no details of the testing methodology, just results.
No hardware. No justification of the games.
I love gaming on Arch. But I am stupid enough to use an nvidia card, which means that I have no Wayland. I am eating shit due to the drivers every few months. There are a whole host of good games with awful DRM that I can’t play. Steam is good but it’s still DRM.
The summary provided article gives the vibe that it wants to pump up Linux’s tyres rather than either demonstrating Linux’s superiority or making the case that you should be gaming on Linux.
> Windows 11: All updates were installed and version 23.11.1 was used for the graphics driver. The global AMD graphics profile was set to standard. This means that Radeon Super Resolution, Anti-Lag and Boost were disabled in the test for better comparison with Linux.
> The gaming PC used for the test was a system with AMD Ryzen 7 5800X on an Asus ROG Strix B550-A Gaming (Bios: 3202). The RAM was 32 GB DDR4-3600-RAM (CL18-22-22-44). The CPU was kept at temperature by a Scythe Mugen 5 while the system was in an open structure. The graphics card is a Sapphire RX 6700 XT Nitro+ with active Quiet BIOS. The monitor supported Full HD at 144 Hz. FreeSync was not active.
> A handful of current or currently very popular (Cyberpunk 2077) Windows games were used for the test. This shows the importance of current program packages for performance. The performance under Linux can also be compared with Windows in current titles.
> Since all games are versions designed for Windows, they ran using the Proton translation layer under Linux – specifically Proton Experimental as of 15 November 2023. Tested was carried out in Full HD.
> Finally, the question arises as to whether these results can also be transferred to a Linux system with GeForce? Yes, all titles were able to run; the editorial team had already tested that. However, the performance remained out of consideration for the time being. In the near future, ComputerBase will also address this topic.
Last time Linux + Wine benchmarking was discussed the major caveat was that some features in Wine implemented with empty stubs (e.g. the game may think soft shadows are computed but stub just always returns solid color). So take this methodology with a heavy grain of salt.
You know how in real life the edge of a shadow is usually kind of fuzzy? "soft shadows" is the graphics technique to make in-game shadows fuzzy instead of hard-edged.
I believe it’s a hard shadow with a three sigma Gaussian blur and smoothstep black-to-alpha with a one-half or three-quarter maximum, pre-multiplied to output.
I'm on Nvidia (2070 super), Arch, X11... It's been literally years since I've had to worry about drivers. Install the proprietary drivers (follow the arch wiki) and they "just work" in my experience. I'm not trying to say you couldn't have had issues, but I am saying I'd caution people from seeing that and assuming that they will have issues if they try this.
Yeah same, wanted to write a similar comment. I have played Diablo 4 , Cities, a ton of emulators with the proprietary nvidia driver and never had an issue. Update works 95% of the time, sometimes I have to do a manual 'mkinitcpio' but otherwise it's all working
But for some reason the new open nvidia driver didn’t like my set up yesterday. mkinitcpio log shows it used the hook properly and then I did it manually. Proprietary one works fine and I am chill with that.
But to reiterate, it’s just the the summary doesn’t go into enough detail. It’s just pumping up the tyres.
My comment is not a criticism or evaluation of gaming on Arch or Linux or nvidia. It’s about the article.
I haven't used Wayland in years and have no real intention of trying to again until x11 becomes unusable so I can't answer that.
My initial interactions with waylands, years ago, including attempting to write a compositor from scratch, left a very source taste in my mouth and the strong impression that wayland was going in the wrong direction. Some things have no doubt changed by now, and a lot of cost has been sunk into it, so I'm deleting my rehashing of some of those concerns as no longer relevant, but I just have no interest in being an early adopter of it.
Steam is the only platform that provides a Linux client, and Valve has provided massive support to Linux with the development of Proton, SteamOS and the Steam Deck.
I have bought a few games in GOG, but until the give us proper Linux support, I won't give them any money.
> until the give us proper Linux support, I won't give them any money
Technically, they don't need to do that - you can run GoG games via Heroic Launcher or even via Steam+Proton by adding them as external games.
If your stance is of a more political nature, that doesn't matter.
Personally, I find myself somewhere in between. On the one hand, Steam is really helping improve support for gaming on Linux (and they've made such good progress that I don't even bother with Windows at all). On the other hand, GoG are fully committed to not selling DRM-ridden crap.
I know that there are alternative launchers, I play the 6 GoG games I have through Lutris. But if I pay for a product I don't want to mess with unofficial and unsupported third-party launchers, at that point I might as well pirate the game.
I'm not sure I follow. The "launcher" really is the .exe . This is especially true for any DRM-free game or application, which we have already agreed is all of those bought on GoG.
WINE (and maybe Proton) is your compatibility layer because you're trying to run _an application_ (games included) on a platform it's not written for.
The reason Steam has a Linux client with Proton integration is not so you, random Linux gamer, can have an easy time running games on Linux. It's mostly 2fold:
1. so that you, random Linux gamer, can run Steam DRM games on Linux at all;
2. so that they can launch products like the Steam Deck without paying license fees, or having to deal with, Microsoft and Windows.
GoG has no such needs. You already have everything you need to run their games. That is their differentiation. One could argue that this is a better proposition for you, the consumer, than that of Steam.
That is exactly my point. Having to manually download and run the installers, plus keeping up with updates and patches, was the standard 20 years ago, but it doesn't cut it anymore. Plus, if you play in Linux, now you need to fight with WINE and Proton, which maybe be easy or might be a nightmare. In total, how much time do I need to waste to play a game?
On the other hand, with Steam when I get a new game I just click on install, I go my merry way, and less than an hour later the game is ready to play, without me having to bother to configure anything.
> But I am stupid enough to use an nvidia card, which means that I have no Wayland. I am eating shit due to the drivers every few months.
Funny you would say this; I'm on AMD (Ryzen 6950HS IGD) and I'm eating shit due to the (FOSS) drivers every few days… to be fair I have two choices of distinct shit with X11 and Wayland, but neither is actually stable.
And none of this affects gaming, just the desktop in general :/. The RX 6950 in the desktop is in fact working without issues (same driver stack). I suspect it's some memory management problem with the CPU/GPU sharing on the 6950HS. But I have no idea how to track it down…
> You don't describe the problems with anywhere near enough detail to guess at what the underlying issue might be.
That's because this is the HN comment section, not the f.d.o bug tracker.
Since you seem to care, what I'm seeing is likely #2220, though the real problem for me is that dmesg output is not particularly helpful on its own and it's the entire desktop that crashes (as opposed to a single application that you could set Mesa/amdgpu debug flags on.) The fact that it "only" crashes about once per week complicates things further; aimlessly enabling debug options would probably result in getting buried in logs.
[drm:amdgpu_job_timedout [amdgpu]] *ERROR* ring sdma0 timeout, signaled seq=24191109, emitted seq=24191111
[drm:amdgpu_job_timedout [amdgpu]] *ERROR* Process information: process pid 0 thread pid 0
This is really kind of a debugging worst case. There's pretty much no useful information here.
Thanks but I wasn't asking for suggestions, forcing to a high power mode is not helpful on a laptop, and various different values for ppfeaturemask have been discussed in the fd.o bug (none of which helped tackle the issue).
I don't expect it to be the cause, but who knows: does your GPU happen to be sitted in the bottom-most PCIe slot of the motherboard ? I encountered weird stuff with the PCIe lines connected to the motherboard chipset (instead of those connected directly to the CPU) on several B550. Some PCIe packets would apparently be garbled/lost when under load, resulting in the driver for the associated peripheral timing out.
It's an OSS game launcher that takes the place of Steam, and you can set things up to run locally so you don't even need an account on their system (lutris.net).
there's also Goldberg Steam emulator that's effectively a 1 line config for many Steam games, but if the exe uses DRM you still have to unwrap it or crack it:
An alternative (which I use) is GPU passthrough, which works on laptops that have a physical HDMI output to the Nvidia card.
I'm only a causal gamer, but never had any compatibility by simple running all games in a windows VM. This setup also serves as a way to sandbox all the windows spyware
I am personally reaping the benefits of their labor. I do wonder how much business impact it has. The steam deck has sold about 1/10 of the units as the PS5. On Steam all Linux is 1.5% of users.
I think the biggest business impact might be behind-the-scenes negotiations with Microsoft.
Of course, this is a seed that might bare more profitable fruit in the future.
I am not sure why you compare PS5 sales to Steam Deck, Playstation is a brand that has been around from the 90s, has been advertised left and right, and so is known in every angle of the world, and are not even the same format/segment, I'd compare Steam Deck sales to ASUS Rog Ally if anything, since they've been around for pretty much the same and have been advertised as much?
They did recently release "PlayStation Portal" which has the form factor of a portable gaming unit, but its not really a portable gaming console, its a portable hardware streaming client for PS5 games.
They would have an edge if they would actually be pushing for native GNU/Linux games, as it is, it is just like OS/2 runs Windows better than Windows, kind of thing.
Just wait until the other SteamDeck competitors get more market share, all of them running Windows.
> They would have an edge if they would actually be pushing for native GNU/Linux games, as it is
The game using a PE or ELF for its binary is irrelevant, what matters is the game actually running. There was a fringe loud minority that went all "No Tux No Bux" a few years ago but nobody really cares about that anymore.
> it is just like OS/2 runs Windows better than Windows, kind of thing.
This is a myth, OS/2 was only able to run 16bit Windows software and that was when 32bit Windows software - which OS/2 was unable to run - was starting to become relevant. In addition OS/2 wasn't even able to run 16bit Windows software as good as real Windows as there were various small incompatibilities (e.g. try running Delphi 1 on it and see all the visual gliches). And finally OS/2's (in)compatibility with Windows was among the smaller of its issues - more important was the lack of hardware support at a time when 3rd party "PC compatible" hardware was the majority of PCs and lack of support by IBM themselves (the story mentioning how IBM saw OS/2 as an accessory to their mainframe business has been posted a few times over the years here).
And let's not forget the anticompetitive licensing scheme Microsoft had for their OSes that practically forced people to get DOS/Windows unless they explicitly asked for something different (that would end up more expensive).
I use POP OS lately, and as much as I want it to succeed, it is unstable when running games. Civ6, random crashes. TF2? Random crashes.
I really dislike Microsoft, but I'm mid-30s, and my whole life, the gaming support for Linux has been tangential at best. I want to convert to Linux fully, but I can't commit yet. It's still too unstable. I have a feeling I will die unhappy and unsatisfied with Microsoft's marketing's answer to graphics called DirectX.
This is probably related to your hardware. I've been using Steam to play games -- Civ6, BG3, Pillars of Eternity -- and everything has worked fine across multiple computers and the Steam Deck. A big caveat here is that I've been using AMD everything, due to their Linux support being better, and also doing everything through Steam, because Valve makes stuff work well.
I think that pre-Proton, Linux gaming was somewhat marginal -- some things worked, but a lot of AAA stuff didn't -- but these days, it's much, much better.
I basically never run the native Linux versions anymore, Proton is always more stable and more performant.
I think that's actually Valve recommendations to developers as well: don't bother with the fast-moving and always breaking APIs and ABIs of native Linux support, just make sure your game works well with Proton.
It might just be personal experience but I've found things always look crappier on Linux than on Windows. I have been off-and-on Windows and several Linux distros on my laptop, and I have found GUIs and multimedia graphics on Windows are always a tad bit more vibrant, clearer and better-looking in general than on Linux. There is just something inexplicably wrong with how Linux setups render things, even in the OS's own GUI (for both KDE and GTK-based DEs) - fractional scaling (if it exists) just makes everything blurrier and worse-looking, fonts look wacky with uneven kerning and ineffective anti-aliasing, colors are dull, etc.
Not to mention GPUs just go ballistic and drown out the game's sound with fan noise every time even when you're running a simple 2D game. You always have to keep the iGPU on and the discrete card off by default because it apparently has just two modes - off, or full-power. Smart throttling like on Windows is non-existent.
I don't mind the extra effort needed to set things up and gather knowledge on running it for the trade-off of a truly free system. But these things are out of any user's control and keep me from completely shifting over.
What ICC color profile do you use? do you use a driver setting like NVIDIA digital vibrance? what DPIs are you using?
Are you using font antialiasing? if so, what font hinting setting are you using? are you using subpixel ordering?
What CPU frequency scaling governor setting are you using? What power setting did you set for your GPU? NVIDIA has Powermizer settings which have adaptive clocking for the GPU and there's a GUI control panel for it that comes with the driver.
Most of what you've mentioned here sounds like you misconfigured your system (e.g.: wacky fonts). Default settings often do not produce the problems you've mentioned.
If you have a HiDPI setup like a 4K screen, simply set your resolution to 4k (3840 x 2160), scaling to 2x, 96 DPI and get a proper font of your liking. You can even get the fonts used on Windows if you install the TrueType font file (I think "Segoe UI Variable Regular" is the font name you would be looking for), with anti-aliasing y hinting set to Slight.
I didn't configure (or misconfigure) any of what you've listed; I don't even know how to do that yet. All was default settings, both on Linux and on Windows.
The only configuration I did was install Nvidia's correct proprietary drivers and turn on fractional scaling and font antialiasing (with correct subpixel ordering) on Linux. On Windows, even that was on by default.
I find the most brutal part about gaming on linux is not the performance once I'm in the game, but getting into the actual game. Launchers via Lutris (e.g. battle.net) take forever to bootup, which was not the case on windows.
Overall experience feels a little janky, but once you get into the game, it seems fine.
Launching those games launches the launcher. Which then either chain-launch the game, wait for launch button to be clicked, or require a Steam linked login, or combinations thereof. Often the game is the only game the launcher can launch.
There must be business reasons it makes sense but experience wise it's a pointless friction.
No, startup times are still significantly higher which I don't really understand, there's no compilation or binary translation going on but still.
And recently, again, any game that bundles the EA app was a pain in the butt to get running, you had to switch back to very old proton versions and disable dxvk and other shenanigans for the EA installer to work. Once it installed properly you could switch everything back to default, but startup time is still abysmal.
Not specifically, but you should get more familiar with how to use Wine yourself, how to create Wine prefixes, what environment variables to set, how to enable dxvk/vkd3d-proton and such. It's not hard to learn, then you won't need to depend on any of these launch managers.
I'm not using Lutris, just my own scripts for launching both native and Wine games.
And you suppose lutris adds random bloat to make things launch much slower than they actually would, because they decided that's what people would expect from a game launcher? Like, what exactly would that be? Mind you, this is explicitly about game startup time, not lutris itself.
I started out doing what you suggest here, but it's a ducking pain in the butt after a while, every couple months there is a new version of X, then Y is deprecated and you should switch to Z... I do this kind of crap as my day job already, no need to keep going in the evening when I just want to have some fun with friends.
Not sure what the reasons above are for Lutris being slow but they are not game launching as that post mentions.
I prefer to control it myself. The way it's set up - I have a common launcher script and each game has a config file with a bunch of variables.
Haven't needed to update those config files for games much after they are set up, so not sure what your particular X or Y issues are, especially something that happens every couple of months.
Exactly. Gaming performance is not a good reason to switch to Linux. But there are many other good reasons (for instance no ads and telemetry, and no continuous nagging to use Edge).
Being able to get a respectable gaming experience will for many be the reason that they can stay in the OS that they prefer for other reasons.
A solid 20% of games in my steam library have Linux versions and many (tens of them!) were built for Linux first. Its way more common for indie/smaller dev strategy games though. Also have you seen the android gaming market cap? Depending who you trust most numbers I've seen have double the number of players on Android and iOS then Windows, Mac, and Linux combined and 3-5x the revenue.
I was saying for games I own. Not gamers I know. Either way Linux and MacOS have been growing thanks to the increased compatibility through proton, wine, and whiskey and when has increased compatibility ever been bad for the consumer? If I was to play a game or do my taxes why should I be forced to buy a Windows computer is I own a Mac? If I want to write programs or run LLMs why should I be forced to struggle through some obscure arcane Linux distro if I have a perfectly functional windows machine?
And I was talking about how little relevance of Steam on Linux has for worldwide share of gamers, let alone game consoles and mobile devices, being forced to emulate Windows games, and Valve's lack of success in making GNU/Linux native games, as proven by Proton being a thing, and SteamOS Machines market failure.
Proton will never stop being a requirement for the large ecosystem of Valve's gaming library.
It is crumbling clearly, because idea of lock-in was to prevent using non Windows systems to run games. And it's not working out due to technology doing wonders translating lock-in graphics APIs into open ones with good performance.
Using APIs for lock-in is an old anti-competitive trick, so it's fun to see it failing for gaming.
Windows game developers couldn't care less, Valve takes over all the porting costs, while Windows
focused game studios get all the money, with increased sales, without caring one second about GNU/Linux.
Lock-in was MS's game here, not game developers'. MS used game developers' reliance on DirectX to make getting games outside of MS systems harder / more costly due to the need for extra porting. Does it make it more clear? So what Valve did is a smart knight move that breaks MS lock-in.
As long as it emulates Windows and DirectX, they have achieved nothing, specially when SteamDeck clones running Windows 11 take over, as expected during 2024.
Now had they managed to convince said studios to write native GNU/Linux games, or to port their Android/NDK games into GNU/Linux, I would have agreed Valve managed something.
> I love gaming on Arch. But I am stupid enough to use an nvidia card, which means that I have no Wayland. I am eating shit due to the drivers every few months. There are a whole host of good games with awful DRM that I can’t play. Steam is good but it’s still DRM.
Not sure what you're doing wrong, but using arch with latest kernel, nvidia drivers, and I play Steam game daily on it with xorg or wayland to mostly stop using my PS4/5 again. I tend to flip back and forth between xorg and wayland, and xorg I still get odd kde compositing bugs, and wayland the compositor has a persistent memory leak that'll use all 128gb of my ram within a few days. I get lots more fps with wayland (particularly yuzu for switch emulation), but I have to restart kde every few days, which is a KDE problem, not Steam.
Otherwise gaming has been great since Proton picked up, but if you do still have problems, use install Proton GE from AUR. If anything doesn't work with default, it does almost certainly with GE.
If games use invasive DRM, or like Epic/Ubisoft/EA with their garbage launchers, F-em - they don't get my money anyways. Plenty out there that don't to have to deal with crappy mega company BS.
for me, it was anyways the case because of windows 11 limitations (inability to undervolt on laptops) and the hardware issues of my laptop (heavy thermal throttling)
it was much easier in the past (when hyper-v is disabled) with windows compared to using specific tools in linux. however, now it is a matter of whether it is possible at all or not.
on the flip side, the metrics don't matter if the game doesn't run to begin with. many of the online multiplayers are plagued with AC not working on linux (at least the last time i checked).
They tested games through compatibility layers that nowadays work for the vast majority of games, unless the anticheat specifically stops it, and Linux with a compatibility layer was still faster than Windows.
That's because it's not actually an (extra) compatibility layer. It's literally just an open-source implementation of the APIs/ABIs used to support gaming on Windows.
While the implementation is different and may do things differently, the code doesn't fundamentally do more stuff.
That's still a compatibility layer, though, because the windows API is implemented on top of the Linux/UNIX APIs/ABIs. If it wasn't a compatibility layer it would run on any OS.
The code fundamentally does do more stuff because anything strictly OS related (by the standards of the host OS) has an additional layer of indirection. The fact that it's still faster is truly impressive.
I do flight simming in Ubuntu and have zero Problems with my peripherals. I also have some USB devices which work out of the box whereas in Windows I have to install some crappy driver first.
Not saying your stuff is well supported but if it isn't, a bit of work/code from your side could share benefits with all others.
I agree, in fact. An ecosystem this varied is the best protection against the current monoculture. It requires software writers to write against a standard instead of 1 specific implementation that might not be viable next year.
All sign point to windows 11 going insane. Still, you can't leave them. They're a prime example of too big to fail. I'd love to see 700 win32 API implementations.
I've used Linux since the 90s and I'm a huge advocate for it, but I just can't take these tests seriously.
Phoronix has also posted a lot of these tests where Linux seemingly performs better than Windows.
I don't have the technical wherewithal to challenge their methodology but it just doesn't seem right that you would add overhead and still outperform Windows on a system barely used by 1% of gamers. It seems like Linux fanatics fantasizing.
I love what Valve have done, thanks to them I am never bored on Linux and never wish I had a windows PC anymore. But something is rotten in the state of denmark.
It's a common misconception that Proton adds any overhead. It's just an alternative implementation. In some aspects it will be faster (and in some other probably slower) because the Linux kernel does things differently from the NT kernel.
I have not run any benchmark but I exclusive use Linux as my gaming platform. I don't play any multiplayer though and I think that's one aspect where Linux is behind. But for single-player stuff it's perfectly fine.
One thing though is that I have to filter out which games I buy by checking on ProtonDB their level of compatibility. But I can play major AAA games just fine.
There's no emulation, just alternative implementation (kernel, drivers, userspace, including Windows APIs with Wine). I don't play and I haven't looked at the results but it does not surprise me that an alternative implementation can perform better.
Dxvk/vkd3d, although as seen with arc newer cards might actually implement older dx Standards in such a crappy way that software translation to Vulkan is still faster.
Yes, and dxvk was actually available for windows before that and people figuring out that even on AMD and Nvidia you can get notable performance boosts in certain games. Intel just did the sane thing of giving up completely since their implementation was never there in the first place, and from what I heard even the hardware not designed with how those older APIs work, which might give them an advantage over the other two.
No hardware. No justification of the games.
I love gaming on Arch. But I am stupid enough to use an nvidia card, which means that I have no Wayland. I am eating shit due to the drivers every few months. There are a whole host of good games with awful DRM that I can’t play. Steam is good but it’s still DRM.
The summary provided article gives the vibe that it wants to pump up Linux’s tyres rather than either demonstrating Linux’s superiority or making the case that you should be gaming on Linux.