According to the article the only significant FPS gains are in the Performance mode. In Silent mode Bazzite has two FPS more in KCD2, and… one less in Hogwarts.
In Turbo mode, the difference is 5 FPS.
And even in Performance it’s not like you’re running at 13 FPS by default - it’s 47 in KCD2 and 50 in HL.
They even provide a nice summary themselves - on average, Bazzite gains 6.6 FPS.
What do you get in return?
Not stability:
I have 512gb model and am running on low with FSR2 on balanced. The game keeps crashing during every play sessions
it has a xbox service that installs from the windows store, it is hella finnicky, ive had to do a full pc restore in the past to get it to work, no other option online helped
the main thing you get back is a better cpu governor to manage power consumption on 2d games. Reviews like the Phawks points out that microsoft kinda handed it off to the handheld makers to optimize for battery life. So in the instances such as getting 8hours of battery life running dead cells because the system doesnt really need to push that much to run the game, the Windows handheld is stuck on a higher performance clock and has a significantly shorter battery life time.
This would be extremely visible if more lighter games are tested, which typically aren’t for reviews like this because its not really fun to show a bunch of games all hitting 60 if you cap framerate.
the main thing you get back is a better cpu governor to manage power consumption on 2d games
Power consumption is handled by the power mode on the device. You get three modes - Silent (13W), Performance (17W), and Turbo (35W). You can switch between them at will, it’s not controlled by the game.
but a game may not necessarily need 13W, thats the point. youre using up more power than what is necessary because the CPU governor doesn’t know if its necessary or not to actually use all 13W.
The steam deck is controlled the same way(has wattage targets), but it understands when its being underused, to use even less resources to maximize battery life.
If you have a car that has gears, where gear 1 is 25, gear 2 is 50 (old cars), it doesn’t mean there aren’t usecases where you want to go 5/10. its unnecessary to always have to hit said target.
The article said the frame rates are more stable under bazzite with windows fluctuating a lot more even at low power settings
I think that, as long as the framerate doesn’t dip below 30, a regular human won’t notice fluctuating FPS. It’s something you’ll see in benchmarks, but not in real life. For example: a friend of mine was playing Elden Ring and and Halo on the ROG Ally X and didn’t notice any issues with stability or framerate. And he’s pretty anal about this kind of stuff.
I read through that thread you linked and I didn’t see bazzite mentioned anywhere.
True, it was about Steam Deck performance. But I had similar experiences on Garuda Linux on my PC with Hogwarts. Every now and again it would just crash and burn. Got better after a Proton patch or two, but that’s the problem with Linux gaming - you never really know what you’ll get. With Windows, you don’t have that issue at all.
You can complain about click bait all day, that’s fine, but you came in here misinterpreting the article and then posted a non-relevant link to support your argument, followed by another back-pedaling comment to say “frame rates don’t matter as long as it’s over 30fps” LMAO.
I think you’re arguing just to argue, seems a little more childish than anything I said honestly.
You definetly do notice the fluctuating fps if it’s more than 10 fps at these frame rates. It feels like it’s speeding up then slowing down over and over. That’s why you should just lock your fps to something the device can maintain.
You will notice “average” 30 FPS plenty when it fluctuates. If it STAYS above 30FPS the whole time then sure but if 30 is the average that shit is struggling and you’ll be getting like 20-40 FPS
They did, yeah. Stupid move, especially just before the Ally release, but shareholders must be appeased.
But, other than the price, it’s been pretty great while I used it. Lots of day one releases, lots of less known titles I would never have tried if not for GP.
Games don’t crash on Windows >because of Windows<. Games do crash on Linux >because of Linux<.
As in: games are inherently compatible with Windows, while on Linux you need Wine/Proton, which is just an extra layer of complication that can cause problems.
I think that, as long as the framerate doesn’t dip below 30, a regular human won’t notice fluctuating FPS.
This is complete bullshit, 30 fps is playable for most games, and I have in the past bumper graphics until fps dip to 30/45 because depending on the game 30 fps on high is a better experience than 60 on low for me. But to say that a regular human won’t notice it is bullshit. There’s a game I play on my deck, for some reason it’s very sensitive to disk usage, so if I’m downloading stuff it dips to 30, and I always have to go and stop the download, because if you’ve been playing at 60, 30 feels very sluggish.
Perhaps you should read what I wrote again, you clearly stated a regular human won’t notice fluctuating FPS as long as it doesn’t dip below 30, and I’m saying that is bullshit, I (and everyone else I know) can definitely see a deep to 30 fps even if it doesn’t go below it.
First of all, read again, no one is talking about below 30. Secondly, yes, you can definitely notice dips even if for a moment, it makes the game feel choppy, or more precisely like a weirdly encoded video that goes slow momentarily and then catches up.
it’s weird, and refreshing, to find someone who actually reads the article and don’t base his comments only on the header, and also provides some additional info.
my opinion isn’t based strictly on the header. because to argue that the default is silent mode, when its very reasonable to assume that the middle tier performance mode is the default usecase, because it has a 80W/h battery when compared to typical other handhelds. Steam deck for example uses a 40W/h battery. 17W, which is seeing the chunk of the performance gain, is a very reasonable target for a 80W/h battery. Because it would roughly be equivalent to a steam deck at 10W on its 40W/h battery.
Your source is from a game from two years ago when it was new; not only does Proton get big updates all the time, but it’s far more mature in general now than it was two years ago. You lose access to Windows store, but Amazon, Epic, and GOG work through Heroic. Maybe EA and Ubisoft are a problem for some people, but those also might work through Lutris. I haven’t shopped with either in over a decade, so I’m not the best candidate to check.
Yeah, I was able to play through most of AC: Odyssey on my deck, thanks to Lutris. I also use Heroic just because I’ve never wanted to install Epic’s launcher.
My source is just an example of gaming experience on Linux.
I run Garuda (Arch-based) which is a distro “for gamers”. The experience is great, I love it. Everything works fine… most of the time.
But, until Proton got a couple of updates, I was just unable to run Mafia. Until Proton got a couple of updates, Hogwart’s would crash randomly. Right now Cyberpunk runs fine… until it crashes during loading sometimes.
That’s the gaming experience of Linux.
All that for a +6.6 FPS average, and no GamePass.
You lose access to Windows store
Windows Store is irrelevant, you lose the XBox (App)-installed GamePass games, which means that you lose access to a tonne of XBox games.
New games pushing technology features will always lag slightly before they make it to Proton. It’s the nature of reverse engineering. On Kubuntu, on AMD, the only crashes I can name are for The Alters, which I knew from Proton DB ahead of time to expect some chop.
The gaming experience on Windows is to get interrupted by updates constantly, less gracefully handle sleep and resume, and sometimes lose control over the game window when popups come up such that you need to be rescued by a keyboard or the touch screen. Those aren’t just my experiences but also captured in the reviews for this very device. What you gain is compatibility with live service games with invasive anti-cheat and Game Pass. For some people that will be enough, but this isn’t even the first handheld gaming device to show a performance delta in Linux’s favor when tested. I don’t think many people are experiencing this stability problem you are, as it doesn’t reflect in many reviews, and a two year old forum post for a game running on technology that moves this fast doesn’t mean that it’s still happening.
The gaming experience on Windows is to get interrupted by updates constantly
gets updates literally on the same day every month with 14 days time before a forced update
complains about getting interrupted by constant updates
Sigh… I’m not even going to comment on this. Can’t fix fundamentalism.
but this isn’t even the first handheld gaming device to show a performance delta in Linux’s favor when tested
Because, overall, Windows 11 is badly optimised and MS already fired all the competent developers, so, yeah, it’s going to happen. But what Windows gives you, is the guarantee that if you have the hardware to handle a game, it will run. Any issues will be on the side of the game, not the OS.
I don’t think many people are experiencing this stability problem you are, as it doesn’t reflect in many reviews
That’s also part of the “Linux gaming experience” - with all the distros flying around, almost nobody will have the same exact experience. Sure, if everyone installs Bazzite on the Ally, it should be a relatively uniform experience, but - again - I fail to see the point in going through all this, losing the (apparently) excellent touch UI and a unified gaming library, just to get… 6 FPS extra. In one power mode.
Your source for stability issues in Hogwarts Legacy is a single user in the Steam community with other users in the same thread not having issues at all. Seeing that Hogwarts Legacy is one of the most played games on Deck (ranked 11th at time of writing ), I think many more people would report issues if crashes were common.
Furthermore, your TONNE of optional game stores is one. I can‘t really think of a game store, besides Microsoft’s, that doesn‘t work on Steam Deck.
These early performance comparisons definitely have limited value for comparing Windows/Linux performance on the device. But I’m sorry to say that your arguments have even less.
I think we should be careful touting linux as more preformant in games than windows because when people switch they wi find that isnt always the case. There are two situations where linux beats windows. CPU intensive games and on systems where the CPU is really bad and sometimes really bad dx12 games.
In every other case the preformance comes down to the gpu drivers which are undeniably better and more tuned on windows.
But this is not bad. I think its still very convincing to be able to say you can get away from windows and switch to linux barely losing any preformance and even in some cases gain it.
Pretty hilarious if after years of being the scourge of Linux and FOSS advocates, complaining how they could never leave Windows because they need it to play, gamers become our greatest allies, switching in droves to get more out of their hardware and games.
Really, this isn’t entirely new, I remember some games were known to run better on Wine than Windows years ago already (Soldier of Fortune comes to mind).
Is that something that can be tweaked on the Windows side? Because if you can’t mess with it on windows, one may argue that the comparison is valid because Linux allows you to tweak those settings
They are supposed to have the same settings, the Linux ones are just wrong and using more power than told It’s supposed to be capped to 17 but pulling over 20 vs 16 on windows. That’s a lot more lower resulting in higher fps but lower battery life.
Cyber dopamine himself says he’s not a benchmark guy as well.
I like him a lot, he’s really passionate and a positive breath of fresh air online, but the guy is surely stoned nearly 100% of the time. No way I’m taking his technical tests at face value.
60-80% better frames on Bazzite for space marine 2 was just too much to not be an error.
Teams actually works just fine. I’m my case installed from the AUR using the electron already present anyways. Zero issues. More specifically zero additional issues compared to Windows.
It’s the first test bed for every developer, which means something like a headset utility is more reliably going to work on Windows. But it’s impressive even that margin is falling.
Imagine seeing Nvidia drop Shadowplay features to push their own beta app improvements, while the Linux imitator for Shadowplay still works simply and fine, and doesn’t even drop for “DRM detected” issues.
Or trying to install/update Epic/Ubisoft games needing to go through another terrible UI upgrade while Heroic and Lutris still look the same.
A year ago, I tried Linux and felt frustrated about some minor UI inconsistencies and fiddling. Recently, I tried again, and it still had stuff to work through, but I was patient for it because now I’m dealing with all that same shit on Windows.
Oh yeah, though to hotkey audio switching I ended up writing my own bash script which was clunky. Curious if anyone better than myself might take charge there.
I’ve been thinking of switching my framework 13 laptop over to bazzite as a toe dip into the Linux world (other than steam deck) if it’s not too bad I might try dual booting my big desktop gaming PC
Honestly, I like Bazzite because it’s very controller and gaming friendly and you won’t be disappointed with it. That said, for a daily driver workstation computer you might want to try Fedora Kinoite which is very similar but focused towards desktop use.
Also it doesn’t hurt to try both as I said they’re very similar! Would love to hear a follow up on your experience.
One caveat. If you have racing wheels or HOTAS you should check if Bazzite supports them. I ran into that issue with my Thrustmaster T300 where the right kernel module isn’t packaged with Bazzite and adding the module to Bazzite… Well, let’s just say it’s easier to reinstall a different OS than it is to add a custom kernel module to Bazzite.
I know. I found that issue when I was looking into how to get my T300 to work and it’s because of that issue that I’m raising the awareness because that issue has been open for over a year and the last maintainer activity there was months ago.
I get that they’re doing it out of their free time and they probably have more important things to do so I’m not faulting them for not being faster with it, but from the end user perspective you’re just going to fiddle your thumbs until something gets done because doing it yourself has the immutable OS getting in the way and it also defeats the purpose of having an immutable OS.
Meanwhile getting the wheel to work on Nobara went, relatively speaking, so smoothly I don’t even remember what I did to get it working.
Compared to Windows. To be clear I’m just basing that on vibes, and I haven’t done any 1:1 testing, but it’s absolutely not any worse than Windows with everything I’ve tried. But also, even if it was slightly worse, the benefit of almost never needing a mouse/keyboard still would make it worth it.
tomshardware.com
Najnowsze