Wayland-only Gnome (continuation of proposal discussion)

[kjetilho (Kjetil Torgrim Homme)]

This change proposal is not removing X11 libraries, so please discuss that issue elsewhere.

[Espionage724]

GNOME doesn’t have a distro, but only devs Wayland. Distros are dropping Xorg because DEs are going Wayland.

It seems chicken and egg , but where exactly did this begin?

---

I need to understand how everyone is ok with the desktop being slower. I was into Linux for performance since F22, and I’m convinced Wayland can’t be as-performant with stuff like libinput → evdev, and have years of experience major GPUs + games.

I can’t accept an inconsistent mouse cursor on a priority timer or -20 FPS on Dota 2 for non-visible Wayland improvements and I’m back into Windows. Who exactly is benefiting from Wayland, and who is supposed to benefit, because it feels like two different groups; any OS can render a desktop nowadays but I’m only aware of 1-2 doing it efficiently.

So where did everyone-into Wayland begin? I want to see the process that led up to now.

[ngompa (Neal Gompa)]

@Espionage724

GNOME doesn’t have a distro, but only devs Wayland. Distros are dropping Xorg because DEs are going Wayland.

They do actually: GNOME OS. Desktops migrating to Wayland-based technology has been a long-time coming, and distributions have been aware of this.

@Espionage724

I need to understand how everyone is ok with the desktop being slower. I was into Linux for performance since F22, and I’m convinced Wayland can’t be as-performant with stuff like libinput → evdev, and have years of experience major GPUs + games.

This is a weird claim since the input stack is identical between X11 and Wayland.

@Espionage724

I can’t accept an inconsistent mouse cursor on a priority timer or -20 FPS on Dota 2 for non-visible Wayland improvements and I’m back into Windows. Who exactly is benefiting from Wayland, and who is supposed to benefit, because it feels like two different groups; any OS can render a desktop nowadays but I’m only aware of 1-2 doing it efficiently.

In general, the benchmarks do not indicate this. To the contrary, it seems Wayland is much better for gaming than X11.

@Espionage724

So where did everyone-into Wayland begin? I want to see the process that led up to now.

This has been a 15+ year thing, with GNOME adopting it 10 years ago. Feel free to trawl through history and use Wikipedia as starting point.

[leigh123linux (Leigh Scott)]

@ngompa

In general, the benchmarks do not indicate this. To the contrary, it seems Wayland is much better for gaming than X11.

The difference between xorg and wayland in those benchmarks is minimal .

For those that were hypothesizing last week that the lighter-weight Xfce and LXQt desktops atop an X11 session would yield better performance than GNOME or KDE on Wayland, that wasn’t the case with this round of testing. For these Ubuntu 25.04 beta benchmarks while using the Radeon RX 7900 XTX graphics, the packaged KDE/GNOME Wayland sessions were delivering the best gaming performance out-of-the-box on Ubuntu 25.04 beta. Often times it was a tight race between these desktop sessions but there were no instances observed where using an X11-based session provided a measurably better experience.

[ngompa (Neal Gompa)]

@Espionage724

I’m still waiting for an explanation as to GNOME pre-42 and high-Hz mice, that was still real-world questionable up until GNOME 48.1, and still sounds like a gamble on timers for the future.

You have your answer: the issues are specific with Mutter. And it’s possible that the GNOME Mutter developers aren’t much for playing video games where input latency matters, so they did not really prioritize it like KDE Plasma did due to it being important for Valve.

@Espionage724

I real-world test on bare-metal without financial ties. Phoronix couldn’t even get Xorg running a little while back The only benchmark I trust on there is Xonotic’s, and X11 runs better.

I’m also skeptical of the hardware; AMD issues are rampant. Of course they’re not in a position to have Xorg drivers functioning well on new hardware. I have mature Intel iGPU hardware available most laptops and non-K desktops easy.

Honestly, I’m skeptical of you at this point. Intel’s drivers have literally been abandoned for X11 for most of the past decade (it hasn’t even had a release since December 2014!), while AMD only recently stopped prioritizing theirs. Admittedly, this doesn’t matter in the default configuration where we use kernel mode-setting, but then that’s the same as all Wayland environments.

You also are cherry-picking benchmarks to try to make a point that isn’t even relevant to you. Unless Xonotic is the game you play every day, it being an outlier isn’t helpful. And the differences are within the range of imperceptible at ~2FPS, so even if it was the game you play every day, it doesn’t matter whether you pick X11 or Wayland.

Finally, Michael Larabel has been doing benchmarking and actual qualification of Linux systems for more than two decades. If anything, he’s probably the closest to a per-eminent expert in the field that we have. He has far more credibility than you do on this point.

[Espionage724]

(post deleted by author)

[py0xc3 (Chris)]

The original proposal has been processed, and this spin off seems to not follow a common goal to solve anything but just puts forward assumptions against other assumptions to tackle/disprove each other, and it starts again to develop underlying aggressions (and new flags ). It seems to have lost its track anyway, and clearly does not tackle a specific question/problem anylonger. I thus close it.

---

@Espionage724 not related to the “topic closing” reasoning: I know you don’t mean it in a negative or insulting way, but please avoid sarcasm and irony: this can be easily misunderstood among cultures and people of different backgrounds. Especially when you refer to other people or their work, this can quickly (and accidentally) cross the line of our rules.