1. Why not link the original article[1] with the benchmarks and use your favorite way to translate?
2. Curious why they didn't use OG Ubuntu in the benchmarks. Yes Popos(I still don't agree on that name) is just another Ubuntu, but I would have proffered the OG snice that's the most popular worldwide and I don't know what Popos does to make it more "gaming focused" than vanilla Ubuntu.
For another popular Ubuntu alternative they could have also went with Linux Mint. AFAIK both are still default on X11 while latest Ubuntu defaults to Wayland so that could have made a difference.
I think the thing that makes popOS relevant to gaming is how system76 offers a variant with proprietary Nvidia drivers baked in and properly configured to not break with updates.
You have to install that after installing Ubuntu, which isn’t necessarily clear and might be confusing for someone coming from Windows where there is only one Nvidia driver, with the version shipped with Windows being adequate for most users.
If you are gaming on Windows, you are most likely installing GeForce now and updating the Nvidia driver separately from the Windows updater so the install separate driver is probably not foreign or confusing to users.
Also on Mint, the OS installs with the open source driver but upon first boot it shows a notification to allow you to download and enable the closed source driver so it is very upfront and not something that you need to hunt for.
People building their own systems are definitely going to be installing from GeForce Experience, but there’s quite a number of prebuilt users who probably aren’t.
Mint’s approach is definitely better than Ubuntu’s, but pop’s has the advantage of making it clear from the point of download that Nvidia users are covered.
The desktop itself is a weird custom-baked x11/GNOME customization you won't really get on Ubuntu. Not many distributions can claim the title of being truly "unique", but PopOS does deserve the title in my opinion. Even still it wouldn't be my choice for gaming, but I'll defend it's identity as something more than another GNOME/Debian spin.
> Not many distributions can claim the title of being truly "unique"
Sir, I think you've just insulted like dozens of distros out there.
IMHO there are a lot of more "truly unique" distros out there than just spins of Ubuntu with a different DE and scheduler: Nobara(from the benchmark), Q4OS, Antix, Mabox, Linux Lite, Gecko Linux, Spiral Linux, in no particular order, not that they're better than Popos, but they're definitely more unique.
Depends. Personally I don't consider Popos's DE to be a feature but more of a bug. I think the two competing giants, KDE and Gnome, are more solid and see more dev time and effort put into them, and will do just fine to satisfy all users, preferences and use cases, I don't see the need for the Nth+1 DE that fixes problems Gnome & KDE don't have. But that's just my opinion.
From using PopOS, mostly for other reasons like out of the box drive encryption and nvidia support, the DE's purpose as far as I can tell is mostly to provide a reasonably user-friendly tiling window manager without replacing all of Gnome. Though now they are also replacing all of gnome in the next major release, so who knows.
As a sysadmin of machines on several distros, IME Arch is the most "linux-like" in that I'm constantly having to figure out obscure errors to make shit work again after an update.
No snark: I understand that's largely a result of Arch deliberately pushing decisions down to users rather than deciding for you, appreciate that it's a deliberate decision, and am _very_ glad it exists.
I guess if "Linux is about choice" (it's not) then Arch is very Linux-like.
> As a sysadmin of machines on several distros, IME Arch is the most "linux-like" in that I'm constantly having to figure out obscure errors to make shit work again after an update.
This is interesting because it hasn't been my experience at all. IIRC in the last 3 years I only had to fix something after an update once, and that was only because of my own stupidity.
I actually find Arch a lot more stable now than when I was using Ubuntu many many years ago, where something would always break after upgrading to a new version. (But maybe Ubuntu's better in this regards nowadays; I don't know.)
Arch completely broke systemd-boot after a recent update for me on my laptop. After spending half a day trying to figure out how to get it put itself back in to the loader.conf I just put Windows 11 on the laptop. My girlfriend uses the laptop and wouldn't know how to fix this if it happened again. So it still breaks stuff on occasion. Maybe not for you, but it does.
Well, I use GRUB for booting since, let's see... well, since I first installed this system, probably something like ~10 years ago, and it just works, never breaks, and I never have to touch it.
The parent poster was definitely right that Arch pushes all of the decisions down to the users rather than deciding for you, so I suppose how much it breaks depends on how good your decisions are and how much of a Linux expert you are. I could fix virtually any broken Linux system while blindfolded, so I know how to set it up so that it never breaks and doesn't require any maintenance. This probably explains the difference in our experiences.
I tried using Arch for my work machine for a bit. It might be better if I only manually run updates for security stuff, but that's management time I could spend doing my job, not polishing my tools.
It's a terrible benchmark all-together, I don't know why it's getting so much traction. The authors themselves don't understand the differences between these Linux distributions.
The website (computerbase) also seems to host their own downloads of the operating systems they review, very weird.
Nobody else likes to do these kinds of benchmarks because if you do them right the performance differences should be negligible. If not you've probably stumbled upon a bug.
Phoronix actually does Linux benchmarks right. [1]
I was about to post the same thing. Distros the like of Arch are outliers in a landscape dominated by "batteries included" distros, neither of which are more or less Linux.
It's a sentence you would expect from someone with little to no knowledge about Linux.
Metaphysically? Were you trying for philosophically?
Most people I’ve talked to want the freedom that comes with Linux, but also as little pain as possible. Arch isn’t it. There’s a reason Ubuntu became so popular back in the day: it was a free and easy to use into to desktop Linux.
Arch has always felt more like a distribution for people who enjoy tinkering with the distro rather than just using the computer to get things done.
I don't quite understand this benchmark. Why compare distributions? What's really the difference here? Some different X11 or Wayland settings? Different driver versions? Different kernel or kernel settings?
Whatever distribution I currently have, in almost all cases, I should be able to configure those things just as well?
So this is just a test about which distro has the best defaults? Ok, it's maybe interesting to some. But the text is misleading, it writes this as "what is the best distro for gaming". For me, much more interesting would be, what are relevant settings to maybe improve the performance, with my distribution of choice.
I think a lot of you are missing the forest for the trees. The fact that Linux is capable of running Windows binaries as well as Windows does is pretty awesome.
The problem with gaming on Linux I found is that a lot of games won't work. Even if it's advertised to work for Linux, it might not work on your distro, or it might not have controller support or audio. Sure you could try and troubleshoot, but I want to play a game, not work. What if you spend a few hours troubleshooting, you get the game working, and you find that it's not actually fun?
There is ProtonDB (Or WineDB) where you can check supposed working games (at least the emulated ones) and how well they work. Often if they don't work out of the box a fitting patch is posted there and reality is sometimes a older ProtonDB build works better.
However in my experience games just work these days in 9 of 10 cases. And if it doesn't it's the launcher and not the game in the far majority of cases, a issue only found with big shit studios.
I permanently deleted the Blizzard/Battle.net identity I'd had since D2:LoD after dealing with Blizzard Support over Diablo IV.
Hour long waits to get in, frequent crashes and logouts, no actual help or troubleshooting, and my refund was denied for "too much in-game time" within 48 hours of purchase (6 hours of troubleshooting and waiting, about an hour of uninterrupted play). This was on a native Windows install, by the way.
My point is that Linux is a great filter for companies with the worst behaviors. If it happens to be a coincidence that the games that tend not to work on Linux also belong to those companies, so much the better.
Also, if it's not in Wine's AppDB or ProtonDB, try using https://looking-glass.io/ with a Win10 IoT VM that's been "debloated". It's like having Windows as a Service.
I think this is happening less as the Steam Deck forces non-Linux-compatible games (and games rated as Linux-compatible but with outstanding problems) to fix any outstanding issues so they can work properly on the Deck. Lots of games in my Steam list have had updates over the past 6 months making them work on the Deck, and I assume (though haven't tested) that that will also impact their compatibility with non-Deck Linux boxes.
I found that even games with "Linux support" tended to work on some distros but not others.
I wonder if there's a guide to setting up Arch in a way that makes it work as much like the Deck as possible? I'm fairly indifferent to choice of distro, maybe it's worth choosing Arch just for game support.
That's super lame. I wish they made it so SteamOS behaved as much like the deck as possible, so compatibility work for the deck would also transfer to steam machines.
If anything it is taking away the worry of game studios to care about GNU/Linux, beyond ChromeOS/Android/Linux, Playstation/Swift POSIX inspired OSes, just let Valve do the work on Proton and call it a day.
Once one gaming generation doesn't grow up on windows.. It's dead.
When they learn how to do all the things with the in game modding language.
It has no real comeback.
It can open up, become a 2nd Linux, but then why?
I find it interesting that Arch is considered a gaming-focus Linux distro. Is that really the case? From my experience, I've see Arch as a general-purpose bleeding-edge distro.
Both of my main computers are over a decade old. 3rd gen i5, and my laptop runs an Nvidia NVS5400m. Absolutely prehistoric hardware.
Arch is the only distro I've tried that just works out of the box. It's the only distro where Optimus/Prime/Bumblebee actually works and doesn't just run the GPU at full bore all the time.
All of my games work as well as you could hope with this hardware, with no configuration or tinkering up front.
Arch has been the absolute most stable and fast OS I have ever used, bar none. My current installation is the longest I've ever gone without reinstalling the OS, including windows. It's definitely not the most ergonomic OS, but it just works, and it works every time.
Specifically, I use the Garuda flavor. It's advertised as being specifically for games, but I don't know what's special about it other than the garish "gamer" RGB default desktop theme. I'm gonna try Cinnamon next.
I recommend staying away from Arch-Derivatives like Garuda Linux. They really add nothing off value. You only end up having a couple of inexperienced Linux users having messed with your kernel, swap disk, schedulers, governors etc. The first principle of ArchLinux is simplicity and being able to know your system, and these derivatives throw that straight out the door.
Everytime you run into an issue you'll not know if its related to Arch itself or something the Garuda installer has messed with. Which is why people on the ArchLinux forums and IRC aren't eager to help others using a derivative.
Garuda is one of the worst examples with the team being so unprofessional. As an example their emberassingly small wiki contains emoticons. Manjaro already gets shame for not properly informing users of the risks of the AUR, and Garuda absolutely discards security entirely by using a third party repository with pre-compiled AUR packages...
All-in-all you get a whole lot of uncertainty for having some stuff pre-configured. Something a simple install script (that you can actually review and modify) could cover as well.
After using Garuda for a couple of years, I agree completely. If I run into any non-trivial problem, it immediately becomes so much worse because there's no way to know how much of the problem is Arch and how much is Garuda. And only one of those shows up in search queries.
I picked Garuda because my requirements were:
1. Arch
2. Has desktop environment out of the box
It was the first one I found and looked adequately shiny.
I just really dislike raw Arch and having to install my DE from scratch. It always feels like I'm doing something wrong and I'm never sure I configured everything correctly.
It's not specifically, but it's a common suggestion to gamers coming from Windows. As a result there's a self-reinforcing cycle where it becomes better for gaming because people put effort into it, despite that not necessarily being a focus of Arch as a whole
And Ubuntu is Debian based, it's not really Debian though.
I own Steam Deck and I also use Arch as my main desktop distro. While Steam Deck uses Arch as a base, it is slightly tweaked. The file system is read only out of the box and you are supposed to upgrade it together with SteamOS from the Steam settings menu. When connected to monitor and switched to desktop mode, it runs regular KDE desktop. If you want to keep it a vanilla SteamOS (with read only system) but still install apps, you can install flatpaks (for example from KDE application store).
I'm fairly certain that this has been the case in the past. I think i remember when wine was still very unstable some games ran better on linux than on windows. But the caveat was that some of the features were not translated via wine. For example some type of lightning or shading (as in shadows) technique... like you would just not have shadows. So back then it was said that games ran faster because the GPU did less work cause of the missing "features" when doing the translation.
Is this not the case anymore? Does proton translate all of the api calls from say dx12? Were the games tested in vulkan or dx12 and are we sure thy look the same visually?
> But the caveat was that some of the features were not translated via wine. For example some type of lightning or shading (as in shadows) technique... like you would just not have shadows. So back then it was said that games ran faster because the GPU did less work cause of the missing "features" when doing the translation.
Do you have a source or example vid or something to link to showing this? I ask cause I've seen this claim pop up a number of times here and on reddit but never have been able to find concrete examples of it actually happening, in addition to the Wine docs saying (IIRC) that if Wine runs into something that isn't/can't be translated, that the program will crash, not just omit it. The idea that it would just not appear at all seems less probable to me than the game crashing.
Unfortunately no, this is from the memory, from around when I first heard of WINE like 20 years ago. Might have been a discussion on osnews.com but I can't really remember it well.
>The idea that it would just not appear at all seems less probable to me than the game crashing.
Now that you mention it, the story might have been something else... there was news about linux beating windows in gaming, but the issue was not WINE but linux drivers for the GPUs at the time not having some features supported.
Again this is such a random information for me to remember it's all hazy. I'm almost 40 years old >.>
I'm a bit confused, literally any Linux distro has everything in their repositories to install Proton and GPU drivers.
> Arch, which is the most Linux one
I love Arch but it's not the "most Linux" (i dont even know what it's supposed to mean). And for someine trying to make a transition from Windows to Linux it's factually the most hostile.
In w11 I can't set the toolbar to be vertical or to autohide anymore. I couldn't find a single reason they changed this. This has been basic functionallity for decades. It's a truly regressive update.
The encroaching advertising is why I don't boot to 11 any more. Ads in everything from the launcher to office apps. Plus for some reason on this older Acer Nitro laptop Windows 11 spins up the fans like you wouldn't believe. Pop!OS they stay silent.
Everyone is tired of hyper-MBA logic making everything so damned pushy.
Everywhere, it's pushy designs, pushy notifications, pushy popups, pushy integrations, pushy reinstalls of programs you uninstall, pushy and second-guessing about everything you want to do that isn't someone else's plan for your computer.
It's infuriating, and people hate it. But someone out there thinks it's leaving money on the table to not push everyone who will tolerate their pushiness deeper into their dependency. "Gosh this would be so easy if you'd just cave and make a $yetAnother account and your tie your wallet and entire identity to it. Aren't you sick of having to be told that over and over?"
I feel the same about all of big tech right now. A bunch of pushy, self-important asses out to shakedown the fools who don't see them and immediately cross the street. And if you can't tell who I'm talking about, exactly.
How about ads popping up in the system tray and no local user accounts to name two big ones. And my personal pet peeve, can't set taskbar buttons to never combine. It is the same as Windows 10, except less good.
Earth is round, and if you think ai isn’t going to be intelligent then you’re not intelligent yourself. Also don’t use windows daily but atleast I’m not an idiot “m$ bad cos reasons so I can whine on hacker news to my boyfriends in a circle jerk”
A gap of FPS has no meaning. FPS have to be compared with percentage difference. If you want to measure a gap in absolute terms you need to convert the measurements to milliseconds per frame.
The original source features a frametime-based rating as well. Surprise, surprise, Windows 11 and Arch are both identical in terms of performance, while Pop!OS and Nobara are about 9% behind.
I think you're reading or portraying that wrong. Pop!_OS and Nobara are about 9% behind in frametime, i.e. their frametimes are 9% lower. That's a good thing!
The charts don't make this particularly clear, but the text summary mentions "Windows is stronger in terms of 1% percentile frame times.".
It's still open to interpretation what exactly they are measuring with their frame timings, e.g. is Windows consistently generating frames within 10.6ms but Nobora is up to 9% less performant then that, sometimes 11.6ms? Although this only seems to be for the 1% lows. I don't think it's a particularly insightful metric in this particular benchmark.
Not sure what make Pop!_os, which i use daily, more gaming focused than just Ubuntu, which also comes with nvidia drivers... also they tested on amd so even more confused at the distinction.
2. Curious why they didn't use OG Ubuntu in the benchmarks. Yes Popos(I still don't agree on that name) is just another Ubuntu, but I would have proffered the OG snice that's the most popular worldwide and I don't know what Popos does to make it more "gaming focused" than vanilla Ubuntu.
For another popular Ubuntu alternative they could have also went with Linux Mint. AFAIK both are still default on X11 while latest Ubuntu defaults to Wayland so that could have made a difference.
[1] https://www.computerbase.de/2023-12/welche-linux-distributio...