Fedora 24 Alpha X.Org vs. (X)Wayland Gaming Tests

[planetguy]

Why is Wayland slower than X?
X is a legacy architecture with lots of problems, and Wayland is this new, fresh thing that is supposed to solve all problems.
Why is it so much slower?
Shouldn't the performance of Wayland be on par with X.Org or even faster?

The tests are not native X11 vs native Wayland tests, where Wayland's more modern architecture would shine. Instead, the Wayland tests use the XWayland compatibility layer to run X11-native binaries - this means keeping all the overhead of an X server in addition to doing Wayland compositing.

[Rubble Monkey]

The tests are not native X11 vs native Wayland tests, where Wayland's more modern architecture would shine. Instead, the Wayland tests use the XWayland compatibility layer to run X11-native binaries - this means keeping all the overhead of an X server in addition to doing Wayland compositing.

Which, if you ask me, makes it all the more amazing that it has even a fraction of the performance of native X11.

[TheOne]

Which, if you ask me, makes it all the more amazing that it has even a fraction of the performance of native X11.

Is nice and all but if game developers don't port their applications to wayland and you as a gamer want to have the best possible performance you will have to be switching from Wayland to X, which would suck... I will stick with X until Wayland catches up and propagates enough, in the mean time is fun to read Wayland articles but not as exciting as before since it has taken so many years for wayland to catchup to X on the Desktop...

[JohnGalt]

PSA: The experimental mouse pointer locking and confinement protocol is implemented in GNOME 3.20. It only needs support in the toolkits like SDL.

https://github.com/elkrejzi/system-m...ve-mouse.patch

The dev who has been working on the protocol also has his own SDL branch.

https://bitbucket.org/jadahl/sdl/

There is also Xwayland with pointer warp support.
https://github.com/jadahl/xserver/co...d-pointer-warp

I'm curious if Mutter 3.20 has support for properly unredirecting fullscreen Wayland and Xwayland windows.

Nice! I was aware of the libinput work on it, but not that it was so close to integration with gnome 3.20. I'm running gnome 3.20 on gentoo with wayland support (systemd). I'll add those patches to test it out .

[liam]

I do not get what is so fantastic to:
- download 1.5GB file, 7 minutes with 50Mbit/s 4G connection.
- can not switch overlapping windows with a mouse, only by clicking the window if it was visible below the other.
- after selecting some setting not returning to All settings.
- slow user interface action
- could not switch to wayland after logging out, rosesomething live had that feature
- virtualbox hangs when using the Software application. Then I stopped testing and deleted the iso file.

Testing also Debian Xfce Live, it works fast and Synaptic also works in the virtualbox. Testing pc: A8-7600 quad core 3.8Ghz cpu, 8GB 2133MHz ddr3 ram and 240GB ssd.

Now get away from the bridge!

[blackout23]

Is nice and all but if game developers don't port their applications to wayland and you as a gamer want to have the best possible performance you will have to be switching from Wayland to X, which would suck... I will stick with X until Wayland catches up and propagates enough, in the mean time is fun to read Wayland articles but not as exciting as before since it has taken so many years for wayland to catchup to X on the Desktop...

Again it has nothing to do with Xwayland. The only reason is that on Wayland Mutter doesn't unredirect fullscreen windows, yet.
Here you can here it directly from a Mutter maintainer.

https://www.phoronix.com/forums/forum/software/desktop-linux/825305-gnome-3-18-on-fedora-23-x-org-vs-wayland-performance?p=825337#post825337