Comparing Linux gaming distros performance

Jure Repinc@lemmy.ml to Linux@lemmy.ml – 36 points –
Comparing Linux gaming distros performance (with Tuxedo Atlas S)
tilvids.com
11

Did anybody think that they did?

I always assumed they were just easier to set up

Some of them advertised specific performance improvements.

I'm not going to rag on them though. Some of them did have performance improvements and basically created the tools and optimized defaults that propagated to standard distros, allowing the gap to close.

This is the great thing about open source. It benefits everyone. Any good idea that does not have significant drawbacks should get broad adoption. And that's generally how it plays out.

Reputations live on for many years (decades, even) after they are justified.

Garuda advertises a different scheduler so I would think that would make difference. It's also one of the things people recommend to improve gaming performance on Linux. Unfortunately as others have pointed out without 1% lows, there is nothing of value in this video. Saying that with respect to Nick. He should step up his game in this area. Average fps just doesn't tell anything, especially on Linux which is even less consistent than Windows

@somethingsomethingidk @JRepin

Hmmm yes, or at least not worse than a general purpose distro.

I can't find the post now but a user saw that in his computer, Garuda, which is advertised as a Gaming distro, performed worse than OpenSuse Tumbleweed and other distros that I don't remember.

The problem is that there isn't a single benchmark that is complete enough to conclude anything meaningful.

yeah I always looked at it as more of an out of the box thing being the idea. I mean really for all distros to some degree. To some degree people, I recognize there are other things but I can feel the heat of the torches from the pitchfork wielders already even before I post this.

Unfortunately, I can't take this seriously as 1% lows and additional variance due to difference in DE haven't been accounted for.

Furthermore, you bet that Tuxedo OS has done a splendid job at optimizing performance on a device that's sold by Tuxedo. Therefore, I wonder if it's even a fair comparison to begin with.

If you're only testing on one set of hardware, it isn't going to tell the whole story. The results might be very different on an AMD vs Nvidia GPU, or even on a brand-new vs 1-3 generation old GPU.

Probably the most important thing for gaming is driver support and ease of installation. This sometimes runs directly counter to other general-purpose needs.

I'm still on the hunt for a distro where everything I need is easy to install. I don't think any exist, primarily because GPU drivers suuuuuuuck, especially when you need CUDA or ROCm to work.

I’m still on the hunt for a distro where everything I need is easy to install. I don’t think any exist, primarily because GPU drivers suuuuuuuck, especially when you need CUDA or ROCm to work.

Have you looked at the opinionated images by uBlue, i.e. Aurora, Bazzite and Bluefin? FWIW, e.g. for Bazzite, AMD's ROCM OpenCL/HIP run-times are fully supported OOTB and there's a workaround for CUDA. It does a lot of good stuff in general. Heck, I'd argue they're one of the most handsfree and easy experiences you may find on Linux.

Thanks for the recommendation! I was looking at the Fedora family since AMD officially supports RHEL 9. Hadn't gotten as far as to figure out how well that transfers to Fedora and its derivatives. Good to hear that it works.

Yeah it’s true that testing on one machine isn’t really scientific. He should have tested the same distros on his other machines too.

Still I enjoyed the video and I really love his work 👍