Nobara OS, Arch Linux and Pop!_OS beat Windows 11 by a slim margin in fps (delta 8) in Windows native games - Cyberpunk 2077, Forspoken, Starfield and The Talos Principle II. Windows 11 wins in Rachet & Clank.
ComputerBase’s testing was done on an all-AMD test rig, featuring a Ryzen 7 5800X (non-3D) and a Radeon RX 6700 XT.
Update: Windows 11 wins in one game.
Assuming this is the usual case where most games are within noise of each other, the ones that don’t run under linux are excluded, and nobody acknowledges that the need to precache/predownload shaders provides short term benefits.
Its like people miss the good old days of “This is the year of linux gaming. Everything works and is perfect. Okay, those games don’t work. But every game I care about works. Except the ones that don’t”. Like, we really are in a golden age of gaming parity but pretending there isn’t still work to be done serves no benefit.
Yup. Just use the same benchmarks major sites use and note any interesting differences. They usually pick games for specific technical reasons, so most of the work figuring out where Linux is weak is done for you.
I personally play on Linux because I use Linux, but because I think it has better performance than Windows or whatever. That should be the selling point, not slight differences in performance. Show that Linux is largely on par with Windows, and then go through all of the other benefits to using Linux, like privacy, package management, and user choice.
Yeah. More or less the same. Pretty much the entirety of my work day is in a terminal and I have increasingly liked “linux” as a desktop since Mint (and now Plasma) are “more windows than windows” in terms of UI/UX. WSL gets Windows a lot of the way toward the OS I want (a good nix-ish terminal with a strong GUI for day to day), but MS also add more and more spyware and stupidity with every update so…
But holy crap do the evangelists go out of their way to undermine widespread linux adoption. Whether it is pretending that opencad is at all a replacement for fusion 360 or that gimp is comparable to photoshop or it is inflating performance or compatibility numbers.
Like, I’ve tried to switch over a few times over the years. And it has always been a shitshow. ProtonDB goes a long way, but it is also prone to outdated information (since the one person still playing Tribes 2 has no need to try newer versions of wine/proton and so forth). And if you check message boards you get the same skewed bullshit. Which mostly boils down to “Okay, well. I figured out that game X won’t work. And I now assume that these fifty other games I care about won’t either”
These days? it is a lot easier because Valve have put in the work to the point that I can more or less just check games in steam. There is still the risk of a new patch breaking something, but it is a lot closer to the good parts of protondb where the steps to recover to a good build are pretty easy (Armored Core 6 was basically a case of just rolling back a major revision of proton) rather than the shitshow. Which then makes it “Well, game X won’t work. But I am reasonably confident that every other game I care about will run performantly so…”
This is exactly why I don’t recommend my distro, OpenSUSE Tumbleweed. It works well for me, but online help is more limited vs Fedora and the various Debian derivatives. I’ve been Linux only for something like 15 years, and I’d hate for someone to take my advice and have a bad experience.
So I recommend Linux Mint Debian, because I know Debian is solid and Linux Mint has a ton of support. I also tell people to not expect crazy performance and for some games to just not work, that way they’ll be pleasantly surprised when things work better than they expect. As they say, under promise, over deliver.
Of the two main games I play, one doesn’t work on Linux due to the anti-cheat they use, and the other has horrific stuttering while loading game assets.
But Linux works better for the curated selection for this article.