Fedora 21 Gaming Benchmarks: X.Org vs. XWayland To End 2014

[phoronix]

Phoronix: Fedora 21 Gaming Benchmarks: X.Org vs. XWayland To End 2014

As some extra benchmarks over the holidays, here's some fresh numbers when running some OpenGL gaming tests on GNOME running natively on X.Org vs. GNOME on Wayland and the games running through XWayland...

http://www.phoronix.com/vr.php?view=MTg3NTI

[uid313]

Why?

Why is Wayland slower than X?

All we heard about is that X is legacy, bloated, synchronous and has lots overhead.

[dwaleke]

This test is native xorg vs xwayland (not Wayland).

[droste]

And XWayland is X under Wayland which means you have the same problems as X ("legacy, bloated, synchronous and has lots overhead")

[johnc]

Why is Wayland slower than X?

All we heard about is that X is legacy, bloated, synchronous and has lots overhead.

What you heard is mostly marketing crap.

But yeah, running X on Wayland would logically introduce some performance hit, kind of like running D3D on OGL. To counter that, the performance of Wayland would have to be much higher than X.

It's why I've never seen the point of running XWayland or XMir or whatever. Just plain X would perform better and it's readily available.

[Xaero_Vincent]

Yeah and XWayland and XMir are going to be very important for a very long time because few if any of the commerical, proprietary Steam games already released will be reprogrammed to target native Wayland or Mir. The performance of Wayland will have to be improved to counter the loss with the embedded Xserver and the proprietary drivers will need to jump on board with Mir and Wayland support.

Thanks to Ubuntu, we'll probably also need to have a Mir to Wayland compatability layer--such as MirWayland so that proprietary Mir apps designed for Ubuntu can run on Wayland used on all other non-Ubuntu based distributions and then WaylandMir so that proprietary apps designed for other distributions can run on Ubuntu.

I personally will be sticking with X11 on Arch for the long forseeable future, until the performance and missing proprietary driver issues are resolved.

[Ericg]

Why is Wayland slower than X?

All we heard about is that X is legacy, bloated, synchronous and has lots overhead.

This is comparing native xorg vs xwayland, so there's an extra layer of pass through it has to do, thus the penalty.

Xorg IS 30yrs of legacy cruft, bloates , synchronous and in some subsystems some convoluted and utterly tangled that the number of people who -truly- understand it can probably be counted on one hand. If you want a REAL xorg vs Wayland fight then there needs to be some wayland-native games that come out that also support X.

[duby229]

Yeah and XWayland and XMir are going to be very important for a very long time because few if any of the commerical, proprietary Steam games already released will be reprogrammed to target native Wayland or Mir. The performance of Wayland will have to be improved to counter the loss with the embedded Xserver and the proprietary drivers will need to jump on board with Mir and Wayland support.

Thanks to Ubuntu, we'll probably also need to have a Mir to Wayland compatability layer--such as MirWayland so that proprietary Mir apps designed for Ubuntu can run on Wayland used on all other non-Ubuntu based distributions and then WaylandMir so that proprietary apps designed for other distributions can run on Ubuntu.

I personally will be sticking with X11 on Arch for the long forseeable future, until the performance and missing proprietary driver issues are resolved.

I disagree, I sincerely hope that anybody who wastes there time writing mir only apps end up having no place to run it....

[Veerappan]

Yeah and XWayland and XMir are going to be very important for a very long time because few if any of the commerical, proprietary Steam games already released will be reprogrammed to target native Wayland or Mir. The performance of Wayland will have to be improved to counter the loss with the embedded Xserver and the proprietary drivers will need to jump on board with Mir and Wayland support.

Thanks to Ubuntu, we'll probably also need to have a Mir to Wayland compatability layer--such as MirWayland so that proprietary Mir apps designed for Ubuntu can run on Wayland used on all other non-Ubuntu based distributions and then WaylandMir so that proprietary apps designed for other distributions can run on Ubuntu.

I personally will be sticking with X11 on Arch for the long forseeable future, until the performance and missing proprietary driver issues are resolved.

And this is why games should be written against libraries like SDL. SDL already has support for Wayland (SDL 2.0, enabled by default in Arch).

Same thing with applications. As much as possible, write against a toolkit like GTK or Qt. The toolkit can choose whatever back-end it wants to run on, depending on the user's system... That might be X11, Wayland, Mir, or something like Mac/Win. Writing directly against X11's API sounds like a recipe for nightmares.

[erendorn]

What you heard is mostly marketing crap.

But yeah, running X on Wayland would logically introduce some performance hit, kind of like running D3D on OGL. To counter that, the performance of Wayland would have to be much higher than X.

Wayland cannot "counter" the performance hit due to XWayland. At most, it can be transparent in terms of performance.
Of course, native wayland is indeed expected to give better performance than native X.Org.

It's why I've never seen the point of running XWayland or XMir or whatever. Just plain X would perform better and it's readily available.

?? That's nonsense. Pretty much like saying "why should there be ToGL" just because Windows can run D3D only applications.