Games run faster on SteamOS than Windows 11, Ars testing finds

I mean, the best comparison would be to simply install Linux on the Xbox Ally X and run benchmarks against one with the stock OS.

the Xbox team hopefully didn't get limited too much by the Windows team, and can do whatever it takes to debloat the game mode.

the game mode should literally only run the bare minimum that's necessary for the system to function correctly

The game mode is what makes it interesting.
 
I don't think I'm color blind...

3hvqQyd.png


Orange would be the thread title and the buttons on the top-ish-right. The top bar of this graph looks like a light-ish red, and the bar two below that looks slightly darker. Can someone else chime in and let me know if I'm color blind and never knew it?
I'm colorblind and it looks like orange, green, red from top to bottom.
@March Climber, @Closer, and Topher Topher :

I just took an online Ishihara test, and I got 100%. I then took the Farnsworth Munsell 100 Hue Color Blind test, and I got 100%. I then took the Lanthony D-15 (both Saturated and Desaturated), and I got 100% on both of those. Maybe all of your eyes are broken and mine are normal? :messenger_grinning_sweat:

In case anyone wants to try this out (they're quick), here you go:
f9RBL9QsNDyg9btx.jpeg


lol…
 
Obviously. Windows and linux don't and can't use the same drivers. They have drivers written specifically for them, so how exactly do you propose they do a "scientific comparison".

So if the issue is possibly Lenovo's drivers for Windows, why are you waffling on about 'Windows bloat and overhead' being the cause?

Main takeaway here is those Lenvo drivers are terrible lol

I like how it's claimed only BL3 had comparable performance when if you remove the terrible drivers from the equation all but Returnal and Doom were within the same range as BL3 was.

SteamOS is great and all though.

Certainly.

and they say the next Xbox will run windows lol, it will DOA if that happens.

One would like to believe MS would not shit up the drivers situation for their own device, but stranger things have happened.
 
Whichever moron was in charge of color-coding the graphs and decided to pick two slightly different shades of red needs to be punched in the face.
As funny as the argument over colourblindness is, the fact is there's a whole rainbow to choose from and they chose two colours next to each other rather than, say, a nice blue for easy contrast.

I also think a slightly larger separation between each graph wouldn't go amiss, either.
 
EGADS! On my phone it is orange!

sorcery GIF
I was seeing the same as you and check my C2 TV color settings. I love saturated colors so in the past I increased TV's "Game Color Depth" settings. When I lowered the settings the upper "red" turned into "orange".

But I agree nonetheless: Why anyone sane use close colors when doing comparison is beyond me.
 
Last edited:
So if the issue is possibly Lenovo's drivers for Windows, why are you waffling on about 'Windows bloat and overhead' being the cause?



Certainly.



One would like to believe MS would not shit up the drivers situation for their own device, but stranger things have happened.
Because both "lenovos" and "Asus" drivers perform worse and what do you think drivers interface with? Again on other devices like a Steamdeck which has other vendor drivers steamOS performs better in most games. Why do you think that is? The only person waffling is yourself and no I didn't mention bloat, only OS overhead.
 
Last edited:
Top to bottom

Orange
Green
Red

Doesn't mean you are color blind. My wife and I will go toe to toe on blue-green colors. I'll see green. She will see blue. She's wrong, of course

I assume you've punished her for speaking out of place, correct? I find it's the easiest way to fix a woman's colour blindness.
 
Last edited:
@March Climber, @Closer, and Topher Topher :

I just took an online Ishihara test, and I got 100%. I then took the Farnsworth Munsell 100 Hue Color Blind test, and I got 100%. I then took the Lanthony D-15 (both Saturated and Desaturated), and I got 100% on both of those. Maybe all of your eyes are broken and mine are normal? :messenger_grinning_sweat:

In case anyone wants to try this out (they're quick), here you go:
Categorising colours is different for everybody. This is the equivalent of asking somebody if the colour aqua is blue or green. You can see the different shades but you just catagorise it differently.
 
So for handheld console sure, every bit of power matters. What about desktops with powerful CPUs and a lot of memory?
 




For now Nvidia+Linux is not worth it. Even AMD is equivalent at best.

The difference (AMD side) is probably more pronounced on weaker hardware, hence it can be more easily seen on these handhelds or small desktops with similar configuration.

Once you have higher end, modern parts, the system will basically power through the overhead for the most part. You still have to deal with Windows and MS being absolutely mediocre though.

As an example I discovered that after latest updates I have around 7-8 CoPilot buttons and prompts in Office, including one by a god damn cursor. And general enshitification of OS continues unabated. Most of the stuff can be curtailed through LTSC install and some tuning, but 365 version of Office also can't help itself.


wDYiw9yFOtJ8DKCy.jpeg
 
Last edited:
and they say the next Xbox will run windows lol, it will DOA if that happens.
DOA? With the ASUS drivers most are close. I don't see any reason to assume that a gaming version of windows can't do similar to Steam OS.

Oh wait, MS laid off all the people who could have made this happen.
 
I scanned through the article and they don't mention it but like all these "tests" I bet they are running everything at 15w.

But I play my rog ally at 19w with handheld companion locking gpu to 1500mhz and cpu averages 2.5-3ghz with that wattage. I get about 1.5 hours battery life but carry a mini 65w powerbank that gives me another 2 hours battery. with these settings I'm able to play most games 1080p high/medium settings and get from 50fps all the way up to 80fps on most games from last 5 years.

And plugged in 30w I lock gpu to 2000mhz and it's like a generational leap over steam deck.

On a side note I picked up a pair of used xreal 1080p oled xr glasses for £120 and paired with the rog ally it's such a great portable setup it's amazing.
 
So for handheld console sure, every bit of power matters. What about desktops with powerful CPUs and a lot of memory?
Laptops would probably benefit from reduced power and heat too. And that's a sizeable market.

For PC you're probably looking more at perf improvements than power savings tho.
 
Yup, can confirm cachyos runs it better than windows. Nightrein freezes and crashes my display on windows, on cachyos, perfect smooth no crashes
 
I scanned through the article and they don't mention it but like all these "tests" I bet they are running everything at 15w.

But I play my rog ally at 19w with handheld companion locking gpu to 1500mhz and cpu averages 2.5-3ghz with that wattage. I get about 1.5 hours battery life but carry a mini 65w powerbank that gives me another 2 hours battery. with these settings I'm able to play most games 1080p high/medium settings and get from 50fps all the way up to 80fps on most games from last 5 years.

And plugged in 30w I lock gpu to 2000mhz and it's like a generational leap over steam deck.

On a side note I picked up a pair of used xreal 1080p oled xr glasses for £120 and paired with the rog ally it's such a great portable setup it's amazing.
I do see this a lot in Steam Deck vs "insert Z1E handheld here" comparisons. It doesn't invalidate it but I think a lot of people overlook it.

I'm a big Apple guy but Apple does the same thing when they bring benchmarks into play. A Mac Studio is 10x faster then a PC but what it actually means is that it's 10x faster then a competing PC at the same power draw. That doesn't invalidate it but it doesn't draw the whole picture as obviously if you put a M3 Ultra Studio up against a 9950x3d/5090 build the PC will win overall. The difference is that the entire Mac Studio draws less power then that 9950x3d alone.
 
Does the article confirm that the resulting graphics on the screen are identical?
It's the same settings so should be identical graphics. The only difference is one running linux (with drivers written for it specifically) and the other windows (with drivers written for it specifically). There wasn't any notable discrepancies.
 
This is like a month old news my guy.









Yes Windows overhead is harsh on the Legion Go.

There was someone, who shall rename unnamed, on another thread who ran with contrarian data.

But, yes, depending on the games Linux is clearly better.

and it respects your time and privacy.

It's such a pure computing experience, I've never seen anything on the MS side since Windows 2000.
 
12 FPS is a lot for people who want to squish every single drop out of their systems.
Reaching a point in performance when an FPS difference that small is impactful to an average user drifts towards their card being nearly obsolete anyway. If you're part of an enthusiast subculture that likes to stare at an FPS counter and clutch their pearls at any minute FPS drop then I suppose it could be a concern to you. But the conclusion that the average gamer with an NVIDIA card would be so impacted by a switch to Linux due to performance differences between the two operating systems is absurd. If you stuck two average users in a side-by-side test without telling them which was running which, they would have no idea.

If you have a specific anti-cheat enabled game that you want to play, that makes sense to me. If you just can't be assed to run a different OS, sure I get that. But to quote performance metrics that show small differences and act like there's this vast chasm of performance is ridiculous.
 
Oh no, 13 whole FPS difference on NVIDIA :messenger_tears_of_joy:. With performance that bad it almost makes all the Windows spyware and bloat tolerable.

OQe46IUeJpawrFkU.png
Your point? That could be the difference between a locked 60 or a game dropping under 50fps.

But why should somebody switch over to get 20% worse performance? Just to stick it to Microsoft? Sure, you can do that, I'd rather wait until Nvidia performance is improved so I don't downgrade my own experience.
 
Reaching a point in performance when an FPS difference that small is impactful to an average user drifts towards their card being nearly obsolete anyway. If you're part of an enthusiast subculture that likes to stare at an FPS counter and clutch their pearls at any minute FPS drop then I suppose it could be a concern to you. But the conclusion that the average gamer with an NVIDIA card would be so impacted by a switch to Linux due to performance differences between the two operating systems is absurd. If you stuck two average users in a side-by-side test without telling them which was running which, they would have no idea.

If you have a specific anti-cheat enabled game that you want to play, that makes sense to me. If you just can't be assed to run a different OS, sure I get that. But to quote performance metrics that show small differences and act like there's this vast chasm of performance is ridiculous.

People make GPU buying decisions on data such as that. Nothing wrong with it. And you very well may not be able to tell the difference when it is 95 fps vs 82 fps, but what about when performance is 50 fps on one and drops down into the 30s in the other. That's not nothing.

I spent a lot of money on my PC and I paid for better performance. Looking at benchmarks to determine how to best maximize my investment is only natural.
 
Your point? That could be the difference between a locked 60 or a game dropping under 50fps.

But why should somebody switch over to get 20% worse performance? Just to stick it to Microsoft? Sure, you can do that, I'd rather wait until Nvidia performance is improved so I don't downgrade my own experience.
I don't agree with the framing that people are switching just to stick to to Microsoft. I don't have any beef with Microsoft, I just don't want them harvesting my data for their LLM and making top down decisions on how I run hardware I pay for. I generally have the same criticisms of Apple. A game dropping to 50 FPS from 60 FPS is not worth the spying, bloat, and lack of user control that Microsoft pushes on users.
People make GPU buying decisions on data such as that. Nothing wrong with it. And you very well may not be able to tell the difference when it is 95 fps vs 82 fps, but what about when performance is 50 fps on one and drops down into the 30s in the other. That's not nothing.

I spent a lot of money on my PC and I paid for better performance. Looking at benchmarks to determine how to best maximize my investment is only natural.
Personally, If a game is at 50 FPS for me I'm either not playing it or I'm about to upgrade my hardware. In the latter case, being able to sink money that would be spent on a Windows license into a GPU/CPU upgrade would probably yield more of an FPS increase than any Windows vs. Linux FPS difference.
 
Last edited:
I don't agree with the framing that people are switching just to stick to to Microsoft. I don't have any beef with Microsoft, I just don't want them harvesting my data for their LLM and making top down decisions on how I run hardware I pay for. I generally have the same criticisms of Apple. A game dropping to 50 FPS from 60 FPS is not worth the spying, bloat, and lack of user control that Microsoft pushes on users.

Personally, If a game is at 50 FPS for me I'm either not playing it or I'm about to upgrade my hardware. In the latter case, being able to sink money that would be spent on a Windows license into a GPU/CPU upgrade would probably yield more of an FPS increase than any Windows vs. Linux FPS difference.

You're on a gaming forum. The priority here is gaming. Most here dont care about your preference for Linux if it has nothing to do with gaming performance. Linux is worse than Windows in gaming for multiple cited reasons. You're going to have to learn to live with that fact.
 
You're on a gaming forum. The priority here is gaming. Most here dont care about your preference for Linux if it has nothing to do with gaming performance. Linux is worse than Windows in gaming for multiple cited reasons. You're going to have to learn to live with that fact.
It's funny you say that in a thread with the literal title of Games run faster on SteamOS than Windows 11, Ars testing finds.
 
Last edited:
I don't agree with the framing that people are switching just to stick to to Microsoft. I don't have any beef with Microsoft, I just don't want them harvesting my data for their LLM and making top down decisions on how I run hardware I pay for. I generally have the same criticisms of Apple. A game dropping to 50 FPS from 60 FPS is not worth the spying, bloat, and lack of user control that Microsoft pushes on users.

If a game is at 50 FPS for me I'm either not playing it or I'm about to upgrade my hardware. In the latter case, being able to sink money that would be spent on a Windows license into a GPU upgrade would probably yield more of an FPS increase than any Windows vs. Linux FPS difference.

That's fine but everyone is going to make that decision for themselves and they will use data such as this to do so. I'm in the camp of waiting on full Nvidia support before I move to Linux, which I fully intend to do. My gaming rig has one purpose: gaming. Since I get better stability/performance with Windows, that's what I'm using.
 
I don't agree with the framing that people are switching just to stick to to Microsoft. I don't have any beef with Microsoft, I just don't want them harvesting my data for their LLM and making top down decisions on how I run hardware I pay for. I generally have the same criticisms of Apple. A game dropping to 50 FPS from 60 FPS is not worth the spying, bloat, and lack of user control that Microsoft pushes on users.
I'm not a fan of that either. That's why I use a Mac for my personal work and my gaming PC just for gaming. I'll switch over as soon as the Nvidia/Vulkan issues are sorted out, but I'm not giving up some significant performance just until then.
 
It's crazy how hard people are trying to push this Linux shit. The headline and the article have nothing to do with each other, but the Op knew that, as did Ars. A fucking handheld with an AMD gpu runs faster on Linux.

Reaching a point in performance when an FPS difference that small is impactful to an average user drifts towards their card being nearly obsolete anyway. If you're part of an enthusiast subculture that likes to stare at an FPS counter and clutch their pearls at any minute FPS drop then I suppose it could be a concern to you. But the conclusion that the average gamer with an NVIDIA card would be so impacted by a switch to Linux due to performance differences between the two operating systems is absurd. If you stuck two average users in a side-by-side test without telling them which was running which, they would have no idea.

If you have a specific anti-cheat enabled game that you want to play, that makes sense to me. If you just can't be assed to run a different OS, sure I get that. But to quote performance metrics that show small differences and act like there's this vast chasm of performance is ridiculous.

Go benchmark actual games that push the hardware. And it's fucking retarded whenever you show benchmarks proving that the performance is worse the Linux bro fallback is "performance doesn't matter". If a game is 13 frames lower it's lows are even worse, and unless you are a 30 fps console owner, you can literally tell the difference with ease. "Yeah it's worse, but who cares" is not a valid argument.

It's funny you say that in a thread with the literal title of Games run faster on SteamOS than Windows 11, Ars testing finds.
It's funny you can't read the article. See my first point.
 
Am I reading it wrong, or are we comparing a Linux version with native legion s go drivers vs Windows 11 version with no native legion s go drivers on the handheld?
 
Debloated windows doesn't really exist.
But it does.

Because if it's possible to get those extra gains by debloating the OS you already have and know, then that's the easier method. Because debloating is really, really easy.

If you really want to scrap Windows and learn to use Linux from now on, that's great. There are a ton of reasons to do that. But most people aren't willing to do that just so they can get a few more FPS in games, in certain scenarios.
 
It's crazy how hard people are trying to push this Linux shit. The headline and the article have nothing to do with each other, but the Op knew that, as did Ars. A fucking handheld with an AMD gpu runs faster on Linux.



Go benchmark actual games that push the hardware. And it's fucking retarded whenever you show benchmarks proving that the performance is worse the Linux bro fallback is "performance doesn't matter". If a game is 13 frames lower it's lows are even worse, and unless you are a 30 fps console owner, you can literally tell the difference with ease. "Yeah it's worse, but who cares" is not a valid argument.


It's funny you can't read the article. See my first point.
Do you always massively strawman someone's argument to win internet points? I said that the difference in performance between NVIDIA cards in Linux and Windows does not make Linux unusable for NVIDIA users. I did read the article. Here are some fun quotes:

"In all other cases, though, even these updated drivers resulted in benchmark frame rates anywhere from 8 percent to 36 percent lower than those same benchmarks on SteamOS."

"Running SteamOS also means eliminating a lot of operating system overhead that the more generalist Windows uses by default. Microsoft seems aware of this issue for gamers and has recently announced that the upcoming "Xbox Experience for Handheld" will "minimize background activity and defer non-essential tasks" to allow for "more [and] higher framerates" in games." If there are no problems in how Windows runs games then why would Microsoft be tweaking the experience to eliminate "operating system overhead."


Performance superiority in certain games on AMD hardware isn't isolated to just handhelds. It shows up in desktop hardware as well:



yd6JyBpHtwVA0JC4.png
C1aR1KrV6d2FmxE8.png


There are caveats, like better overall highs in Cyberpunk with worse lows (which makes the Windows experience better), and there are games where Windows straight up wins. Like others have pointed out, the NVIDIA situation isn't going to last forever. If those performance gulfs keep decreasing what other argument is there to run Windows versus Linux other than the odd anti-cheat game that won't run?

TCDt1dWtzV9ZNDmh.png
 
Last edited:
Do you always massively strawman someone's argument to win internet points? I said that the difference in performance between NVIDIA cards in Linux and Windows does not make Linux unusable for NVIDIA users. I did read the article. Here are some fun quotes:

"In all other cases, though, even these updated drivers resulted in benchmark frame rates anywhere from 8 percent to 36 percent lower than those same benchmarks on SteamOS."

"Running SteamOS also means eliminating a lot of operating system overhead that the more generalist Windows uses by default. Microsoft seems aware of this issue for gamers and has recently announced that the upcoming "Xbox Experience for Handheld" will "minimize background activity and defer non-essential tasks" to allow for "more [and] higher framerates" in games." If there are no problems in how Windows runs games then why would Microsoft be tweaking the experience to eliminate "operating system overhead."


Performance superiority in certain games on AMD hardware isn't isolated to just handhelds. It shows up in desktop hardware as well:



yd6JyBpHtwVA0JC4.png
C1aR1KrV6d2FmxE8.png


There are caveats, like better overall highs in Cyberpunk with worse lows (which makes the Windows experience better), and there are games where Windows straight up wins. Like others have pointed out, the NVIDIA situation isn't going to last forever. If those performance gulfs keep decreasing what other argument is there to run Windows versus Linux other than the odd anti-cheat game that won't run?

TCDt1dWtzV9ZNDmh.png


You're showing 1080p results on that 9070XT for a reason. I use that video too. We both know why you wont show the higher resolution benchmark. Also show the raytracing results for Cyberpunk. You also cropped out the Nvidia part of the benchmark. Sad.
 
Last edited:
Another thread about this. I think some users who miss the console wars have moved the battle to operating systems. And they're doing all this over 3%
 
I'm not a fan of that either. That's why I use a Mac for my personal work and my gaming PC just for gaming. I'll switch over as soon as the Nvidia/Vulkan issues are sorted out, but I'm not giving up some significant performance just until then.
Yep, I do the same. Mac for pretty much everything other than gaming, messing with AI, and some rendering / 3d modeling stuff.

Once Nvidia drivers get close enough I will switch over.
 
Another thread about this. I think some users who miss the console wars have moved the battle to operating systems. And they're doing all this over 3%

This thread was created in June. This isn't a warring thread for most here. Since PC gamers have a choice of operating system, following the progress of gaming on alternatives is of interest to many of us.
 
This thread was created in June. This isn't a warring thread for most here. Since PC gamers have a choice of operating system, following the progress of gaming on alternatives is of interest to many of us.
Even so, it's another thread. Having choices is nice, but it always devolves into a flame war 😂
 
You mean the one that has a hundred uneccessary things running in the background performs worse? Shocking news
 
Last edited:
You're showing 1080p results on that 9070XT for a reason. I use that video too. We both know why you wont show the higher resolution benchmark. Also show the raytracing results for Cyberpunk. You also cropped out the Nvidia part of the benchmark. Sad.
Every single time people like him do this, because if they show the full picture they come off with egg on their faces. He intentionally didn't show the RT benchmarks for The Witcher 3 or Cyberpunk because it would prove the exact opposite of his claims.

This is a Linux channel, btw:



16-20% lower aggregate. And you know why GoW wasn't added? Because it was broken on one of the Linux distros...lol.

Another Linux channel:



15% + aggregate lower, but go look at SH2 or Cyberpunk, or Witcher 3 with RT. It's a beatdown.

A month later, tested again:



When you have to show mid-range 1080p AMD benchmarks to prove your point, you have have no point. They have to pretend that either Nvidia doesn't exist, or that performance doesn't matter in order to push this platform. It's a joke.


And here is another thing the Linux bros won't tell you, that "90% of games work on Linux" is only a half truth. Go to ProtonDb and look at how many games that "work perfectly" need tweaking or multiple tweaks. Go though all the posts and look at things these games need you to do to play. Or that you have to run lower settings because things like RT shit the bed or crash the game or whatever. I'd love to see a fresh Linux install with the top 100 games. No tweaks, just boot the game and play and see what happens. I guarantee 90% don't work as intended.

Another Linux channel, full of people who use and know how to use Linux:



Look at the percentages of people who know their way around Linux and still have to deal with issues all the time. This isn't even gaming, just the overall usability of the OS in general.



The funniest part about this Linux shit is, if Linux were actually better people like me would switch. No one stays on Windows because we love Microsoft. I buy Nvidia because it's better than AMD hardware. I bought my 4090 because it was better than my 2080. I literally want the best experience I can get, so if Linux offered a better experience I'd switch tomorrow. But it doesn't, and that's why they have to grease these benchmarks with certain hardware, or certain resolutions or certain settings. It's all a smokescreen for the reality that Windows is still just better and that's a fact, not an opinion.
 
Every single time people like him do this, because if they show the full picture they come off with egg on their faces. He intentionally didn't show the RT benchmarks for The Witcher 3 or Cyberpunk because it would prove the exact opposite of his claims.

This is a Linux channel, btw:



16-20% lower aggregate. And you know why GoW wasn't added? Because it was broken on one of the Linux distros...lol.

Another Linux channel:



15% + aggregate lower, but go look at SH2 or Cyberpunk, or Witcher 3 with RT. It's a beatdown.

A month later, tested again:



When you have to show mid-range 1080p AMD benchmarks to prove your point, you have have no point. They have to pretend that either Nvidia doesn't exist, or that performance doesn't matter in order to push this platform. It's a joke.


And here is another thing the Linux bros won't tell you, that "90% of games work on Linux" is only a half truth. Go to ProtonDb and look at how many games that "work perfectly" need tweaking or multiple tweaks. Go though all the posts and look at things these games need you to do to play. Or that you have to run lower settings because things like RT shit the bed or crash the game or whatever. I'd love to see a fresh Linux install with the top 100 games. No tweaks, just boot the game and play and see what happens. I guarantee 90% don't work as intended.

Another Linux channel, full of people who use and know how to use Linux:



Look at the percentages of people who know their way around Linux and still have to deal with issues all the time. This isn't even gaming, just the overall usability of the OS in general.



The funniest part about this Linux shit is, if Linux were actually better people like me would switch. No one stays on Windows because we love Microsoft. I buy Nvidia because it's better than AMD hardware. I bought my 4090 because it was better than my 2080. I literally want the best experience I can get, so if Linux offered a better experience I'd switch tomorrow. But it doesn't, and that's why they have to grease these benchmarks with certain hardware, or certain resolutions or certain settings. It's all a smokescreen for the reality that Windows is still just better and that's a fact, not an opinion.


It's really simple. For low powered fixed hardware products SteamOS offers superior performance to Windows currently. It also offers nice features like shader distribution that the Windows PC handhelds need to catch up with. For desktop variable hardware PC gaming, Windows is superior to Linux in every aspect. Its not even a debate. I dont think people even use SteamOS on Desktop PCs. Even with the handhelds most aren't actually using SteamOS when they make videos.
 
Last edited:
It's really simple. For low powered fixed hardware products SteamOS offers superior performance to Windows currently. It also offers nice features like shader distribution that the Windows PC handhelds need to catch up with. For Desktop PC gaming, Windows is superior to Linux in every aspect. Its not even a debate.
Windows 11 27 H1 preview update available! Install now? H2 preview ready! Reboot to perform changes!

Breaks 100s of things.

Yes, you can just ignore it, but I HATE how the update shit flashes in the taskbar and popups every few days/weeks. SteamOS seems a lot less intrusive for their updates, and on a desktop OS, I would expect to be able to hit uptimes in the years, not weeks, or maybe months, like Windows.
 
Top Bottom