Ubuntu 20.04 Gaming Performance Across Desktops, X.Org vs. Wayland

[Ifr2]

Sorry if I'm inopportune, but today the only thing I can't do with Wayland is screencasting. I use GNU/Linux for all my tasks: programming, virtualisation with VirtualBox, gaming, common desktop tasks (web navigation, music, video, Prime Video, etc) and I feel it very good. I use Wayland on my laptop Acer Aspire A515-54 and DiRT Rally works stabler on XWayland than Xorg (I had a few hangs on Xorg).

Obviously, to see old applications working on XWayland is not ideal, but the day when OBS works properly on Wayland I can fully migrate. I hope with the passage of time to see the applications migrating to Wayland and see XWayland only patched for some old native games, because Wine games could work better if Wine runs natively on Wayland.

Today I use Wayland on my laptop, but on my desktop, where I do screencasting, I still keep Xorg.

The problem with NVIDIA is NVIDIA's fault to no matching its driver with the standards for the Linux graphics stack.

After 8 pages of discussion, I think this is the most sane comment of all.

To me, it doesn't matter if Xorg is doing better or worse than Wayland TODAY, since atm we can choose to use both of them depending on our needs, but eventually and just because there happens to be a general interest into switching to Wayland, Wayland will be the default. In the meantime you choose the hell you rather working with and that's it.

In my case, I use both, Gnome+Wayland on my laptop, Sway at work, Gnome/Plasma + Xorg on my desktop so I can use my Nvidia GPU to play CSGO.

[mppix]

The world doesn't revolve around Gnome/KDE which still can't get Wayland where Xorg already is. It's been almost 12 years! How long until we have basic stuff like desktop recording in a good working shape? I don't even care about Xorg compatibility that much - I want the core stuff to work perfectly!

FTR, I think with Gnome 3.30 (Debian 10 Buster default) Wayland reached about parity with Xorg. Occasional hickups and restarts are needed with both.
As of Gnome 3.36, Wayland is flat out better. No blinking, no tearing, responsive, and so far rock stable (stability needs long periods of evaluation).

There are still things to improve (see original proposal for Gnome 4) and we are still in the transition phase. Some of the benefits will also only become clear down the road when applications pick up the benefits and we can start removing the compatibility interfaces of the old stuff.

[mppix]

That in fact wrong. IPMI/KVM displays on servers are normally GBM. You general framebuffer text on Linux on open software graphics drivers are using GBM.

I meant "No need for *Nvidia* GBM or EGLstreams" since the display is independent of the compute GPU's..
I made a correction to the original post.

[oiaohm]

Correct but Nvidia is dominant in HPC GPU compute with CUDA and their driver is very good. There, driver support is crucial. However, HPC boxes run without desktop and have with IPMI/KVM displays so no need for GBM or EGLstreams.

That in fact wrong. IPMI/KVM displays on servers are normally GBM. You general framebuffer text on Linux on open software graphics drivers are using GBM.

The objective of GDM was to end up with the single thing managing the graphic card buffers be you in text mode or graphical.

EGLstreams still leaves you with the problem where framebuffer can come into dispute with normal graphical output.

[treba]

That's all nice once we have it in a working stable state with no corner cases and pitfalls.

There are lots of corner cases and pitfalls on X11.

And the fact that many DEs under Wayland have to reimplement a lot of very complex stuff remains. The world doesn't revolve around Gnome/KDE which still can't get Wayland where Xorg already is.

Gnome, KDE, compiz etc. have been very complex on X11, too. Especially as they needed to work around X11 shortcomings. In fact they have been compositors for a long time (thus the "fullscreen unredirect" stuff for example). Simple X11 window managers don't offer what most people want from a modern OS.

It's been almost 12 years! How long until we have basic stuff like desktop recording in a good working shape? I don't even care about Xorg compatibility that much - I want the core stuff to work perfectly!

I guess almost everyone agrees that things are not going as fast as they could. But hey, on Gnome 3.36 you now get properly sandboxed and low overhead desktop/window sharing on Wayland (as long as apps use the propper APIs). And by the way - it's not like DE devs stopped working on X11 12 years ago - if they had (and had invested all their time into Wayland) we likely wouldn't have this discussion. But again - what about helping with that instead of calling it shit in some random forum?

[oiaohm]

Seems like you've made birdie's point - far fewer than 1% of GNU/Linux computer users are playing Portal on Wayland. So the 99% number probably stands as an accurate assessment.

No I did not. You did not read though the numbers either did you. Only 2 cases showed any major advantage as birdie pulled out both have the same dependency and it one bug. Generally across the benchmarks Wayland either matches or wins with portal being the biggest win.

So the 99% assessment is not accurate. The correct assessment is 40% on open source support graphics. I have not corrected for the ratio between Nvidia and open source supporting hardawre in use(Amd and Intel) But the number does not land a 99% no matter how you run it.

[horizonbrave]

The fact of the matter is that Xorg still works beautifully for 99% of people out there while Wayland still has teething issues even though it was meant to be THE BEST GRAPHICS SYSTEM IN THE WORLD EVAR. But that will surely change - we just have to wait a few more years. Ten will be enough I guess. Right? 22 years of development must surely be enough!

What Wayland fans fail to realize is that you don't swap something which is working fine for absolute most people out there with something which still has a ton of issues and corner cases and requires special graphics drivers and which doesn't work well or doesn't work at all with NVIDIA.

And to add insult to injury Wayland requires a ton of work/code from the environments which want to work with it that's why XFCE/IceWM developers have simply ignored it so far.

are you a dinosaur or just a fucking nvidia crack head?

[birdie]

I give you a hint who these Wayland fans are: Xorg devs

Don't think there's a way to convince you, but FTR: input (touch, gestures, kinetic scrolling...), security (sandboxing, lock screens...), HiDPI (fractional scaling, multi monitor with different scales...), power efficiency (not updating hidden clients - solved via frame callbacks on Wayland, hardware overlays or similar future technology), clean and versioned APIs, pixel perfect presentation - all areas that are incredible hard to get right on X11 (not just Xorg) and straight forward on Wayland. I can really just ask you to talk to people who actually work in these areas - it has a reason why the devs who know most about it push most for the switch to Wayland

That's all nice once we have it in a working stable state with no corner cases and pitfalls. And the fact that many DEs under Wayland have to reimplement a lot of very complex stuff remains. The world doesn't revolve around Gnome/KDE which still can't get Wayland where Xorg already is. It's been almost 12 years! How long until we have basic stuff like desktop recording in a good working shape? I don't even care about Xorg compatibility that much - I want the core stuff to work perfectly!

[willbprog177]

I'm sad Fluxbox wasn't tested. What is the contact info for the Phoronix "window manager inclusion officer"??

[treba]

Oh, yeah, imaginary Xorg bugs and use cases for 0.1% of users out there. Wayland fans are obsessed with them for the lack of any strong arguments.

I give you a hint who these Wayland fans are: Xorg devs

Don't think there's a way to convince you, but FTR: input (touch, gestures, kinetic scrolling...), security (sandboxing, lock screens...), HiDPI (fractional scaling, multi monitor with different scales...), power efficiency (not updating hidden clients - solved via frame callbacks on Wayland, hardware overlays or similar future technology), clean and versioned APIs, pixel perfect presentation - all areas that are incredible hard to get right on X11 (not just Xorg) and straight forward on Wayland. I can really just ask you to talk to people who actually work in these areas - it has a reason why the devs who know most about it push most for the switch to Wayland