GNOME X.Org vs. Wayland Performance + Power Usage On Fedora 32 With AMD Renoir Laptop

[phoronix]

Phoronix: GNOME X.Org vs. Wayland Performance + Power Usage On Fedora 32 With AMD Renoir Laptop

As part of our ongoing testing of the AMD Ryzen 5 4500U and Ryzen 7 4700U "Renoir" mobile processors, here is some Wayland vs. X.Org data with the GNOME desktop on Fedora Workstation 32...

http://www.phoronix.com/scan.php?page=news_item&px=GNOME-Xorg-Wayland-AMD-Renoir

[pranav]

Does that mean I should stop using Wayland?
Why there is no improvement yet?
Is this mainly because of Firefox not being so great on Wayland?

[kokoko3k]

...but... my X11 is working very well right now!

[s_j_newbury]

Does that mean I should stop using Wayland?
Why there is no improvement yet?
Is this mainly because of Firefox not being so great on Wayland?

Obvious answer is Michael is using Xwayland for all these benchmarks which is an additional layer compared to running directly on Xorg, extra memory usage gives this away. I don't know why many distributions still default to using Xwayland even while defaulting to a Wayland Gnome session, but that doesn't force everybody to do so. Even if your distribution defaults to Xwayland there are environment variables GDK_BACKEND=wayland (for GTK), MOZ_ENABLE_WAYLAND=1 (for Firefox) and SDL_VIDEODRIVER=wayland (for SDL2 based games) which overrides the default. Personally, on my Gentoo systems I always default to native Wayland because I only want to run Xwayland when I really need it.

I would love to see benchmarks of native Wayland vs native Xorg!

[andre30correia]

X11 is dead.

and wayland is using more power, linux need a new display manager and put wayland in garbage

[Sethox]

Does that mean I should stop using Wayland?
Why there is no improvement yet?
Is this mainly because of Firefox not being so great on Wayland?

To be perfectly honest, these tests means nothing. Wayland works as a new development project would, since it's a new implementation there are scopes that need to be covered that has not been seen before (fundamentally). I say this because Wayland is relative new compared to X as a whole.

All we users can do is test, report, bisect and/or wait and see when it becomes better.

[Pepec9124]

But wayland was meant to be more efficient and use hardware acceleration to reduce power consumption.
Maybe in ten more years it will be ready, X11 with stuff like sandboxing is fine.

[blackshard]

Does that mean I should stop using Wayland?
Why there is no improvement yet?
Is this mainly because of Firefox not being so great on Wayland?

Benchmarks just tell one part of the story. Under the cover, wayland should be much easier to be extended and maintained. X.org
also had numerous optimizationa during years take makes it performant and efficient, but as far as I know, it still requires redundant drivers, duplicated code and glue code between 2d and 3d. I think finetuning on wayland will happen when there will be some kind of mass adoption - just think to rhe optimizations received by gnome shell when it become Ubuntu official DE

[tomas]

It's not clear to me: is it testing Firefox wayland against Firefox X11 or is it testing the same Firefox X11 on both X and wayland (through Xwayland) ?

[andyprough]

To be perfectly honest, these tests means nothing. Wayland works as a new development project would, since it's a new implementation there are scopes that need to be covered that has not been seen before (fundamentally).

New development project, that's very funny. It's so old that it's about time that someone started working on a Wayland replacement.