KDE vs. GNOME, X.Org vs. Wayland Radeon Linux Gaming Performance With Ubuntu 18.04 LTS

[aufkrawall]

There are several ways to solve the tearing problem in KDE, I have always used the old method. Create a file in etc / profile.d> create a file with the name tearing.sh> write :
Codice:
export KWIN_TRIPLE_BUFFER=1
It usually works well.

Actually, it never works flawlessly. The only thing that does is Compton with free drivers.

[F.Ultra]

Wayland is a protocol, so it can't be bloatware.

Then you have not seen many protocols. E.g one protocol that comes to mind was the TradElect/Infolect protocols which was constructed by Accenture and Microsoft. The overhead to send just a single byte was huge. That is however not to say that Wayland is bloat (I have no insight in the Wayland protocol) just that a protocol can be bloated.

[schmidtbag]

Most people Claim that for gamers only Gnome or Unity is relevant but kde+wayland bring's the best result for ganers in The Most cases.

I don't know of any gamers who recommend either GNOME or KDE. I'm not saying they should be avoided, but the more lightweight environments sometimes perform better.

I use KDE +gaming for years and every time i See gnome or Unity it feels Like a joke

I don't agree with this. KDE for many years was the worst performer in gaming benchmarks, and only in the past year or so has it managed to keep up. Note, I'm writing this on a KDE setup and I have never had a GNOME setup on one of my PCs, so I'm not biased here.

Wayland is bloatware and a waste of human resources.

If you're suggesting the worse performance of Wayland is because of bloat, you are so very wrong. It's slower because the games have to translate over XWayland. It's kind of the same reason why games almost always run slower in Wine vs native.
There is nothing bloated about Wayland. In fact, it removes bloat - Wayland itself is only roughly 1MB. XWayland is a little over 2MB. Everything else is handled by Mesa or the desktop environments, and that being said, they'll only include whatever is relevant; X11 includes a lot of code collecting dust due to legacy issues. I could go on and on about what makes it less bloated than X11 but it'll probably go right over your head anyway. You so stubbornly want to stick with your old ways that you just can't possibly see how anything different could ever be better or necessary.
I am not saying newer=better, but Wayland involves a lot of things that are much needed for Linux.

[Vistaus]

Most people Claim that for gamers only Gnome or Unity is relevant but kde+wayland bring's the best result for ganers in The Most cases.

I use KDE +gaming for years and every time i See gnome or Unity it feels Like a joke

Correct me if I'm wrong, but aren't most people saying that i3 is relevant for gamers? That's what I read every time gaming is involved or someone reviews i3...

[leipero]

Yea, finally, after four pages someone brought this up. These benchmarks are FUBAR because xwayland is locked to 60fps (and very jittery while gaming). What this means is that PTS is measuring frames that are discarded and never presented on the monitor while running xwayland. Try enabling V-sync and you will see that even on high refresh rate displays the frame-rate under xwayland will drop to 60fps while it should be whatever rate the display is running at.

How about actually paying some attention and maybe try actually playing some of the games instead of putting 100% trust into automated benchmarking? This issue is pretty obvious if one actually plays some games, at least on a high refresh rate display. Try something like CS:GO if you really want to make it super obvious.

https://bugs.freedesktop.org/show_bug.cgi?id=103282

Ok, let's remove some miscnceptions, I don't know if that bug report is valid at all, it may as well be compositor problem, not xwayland. Why xwayland doesn't ignore refresh rate when it was aimed at with "CLUTTER_DEFAULT_FPS=120" (asuming 120Hz) for example? I've reported bug for this on mutter long time ago..., I didn't know what was the case for kwin and others (tho, at that time kwin didn't even support wayland properly), but I have no reason not to trust you on word.

One thing is for sure, input lag on wayland is huge, and the very fact whole protocol was made to suposedly improve latency in rendering (and other security things), It goes beyond me how is that even a thing..., but I do not really care, since X.Org works great for me, there are 0 reasons to change it as far as I'm concerned, and it is far more feature rich, so better in every single aspect.


As for other posts, people complaining about tearing on KDE or whatever, how can one line bother you so much? One horizontal line should be preferable over input lag, and whatever you do to remove tearing it will introduce either input lag or stutter (or both in some cases), isn't mroe responsive input more important than one dumb line that is invisible most of the time?

[Brisse]

leipero What framerate does glxgears report if you run it in GNOME on Wayland with that environment variable set to a non 60hz rate?

I've also suspected compositor issues with GNOME on Wayland but I haven't found any good way to test or prove those cases since I've not been able to find native Wayland software suitable for such purpose.

[leipero]

Brisse I don't know, I can't test it since I am on 60Hz display, and can't re-connect one that is 75Hz because I need to move all the stuff..., and machine there have lubuntu (no wayland). Do "CLUTTER_DEFAULT_FPS=120 vblank_mode=3 glxgears" and try it yourself if you have high refresh rate display (for this case 120Hz). I know that back in Gnome-Shell 3.24.x I've solved stuttering in wine by maching clutter frame rate to the refresh rate of the display, otherwise it would still kick high framerate with vblank disabled, but it would be very jerky/stuttery experience.

[Brisse]

leipero

~$ CLUTTER_DEFAULT_FPS=120 vblank_mode=3 glxgears
ATTENTION: default value of option vblank_mode overridden by environment.
309 frames in 5.0 seconds = 61.725 FPS
300 frames in 5.0 seconds = 59.995 FPS
301 frames in 5.0 seconds = 60.000 FPS
300 frames in 5.0 seconds = 59.998 FPS
300 frames in 5.0 seconds = 60.000 FPS
300 frames in 5.0 seconds = 59.998 FPS

[johnc]

And if Wayland is bloatware what X11 is?

X11 is amazing... bordering on spectacular.

[Danielsan]

I am wondering how many Phoronix readers use Ubuntu as main OS...