Wayland beats the hell out of Xorg in GPU test

[UnkendTech]

this test is done using compositing as there is no compositing bypass for Xwayland as of yet i'm not posting the test on the steam forums but i did post the Link and a Little info on it

"mannerov"
"Since there isn't sync support right now, XWayland doesn't bypass X compositing (whereas the applications could because they are fullscreen).
There is a copy and that explains the lower performance than on bare X.

The way to avoid the copy would be reenabled, when sync to the frame Wayland display is implemented. (and then performance would be the same than on X)."


http://phoronix.com/forums/showthread.php?84236-XWayland-being-under-active-development-again&p=359049#post359049


also they heres a Video of "Embedded weston running on the dedicated card (with PRIME) and XWayland" posted by the same user
http://www.youtube.com/watch?v=gpITLskMkXA&feature

[mannerov]

I think the title is a bit misleading.

If you look closely, the tests show much higher fps for Wayland when the fps of the benchmarks are already very high. It is because the Wayland compositor synchronizes to the screen refresh and won't do more than /refresh-rate/ copies per seconds.
So yes, fps are higher for high fps applications, and only the most recent frame, before the compositor draw the next frame for the screen, will be picked up and displayed.

However, for fps around or lower than 60 fps, X and the Wayland compositor would do about the same number of copies, and performance is similar.

People said than when XWayland would be entirely ready, its performance would be better than X, but it will be true only for non fullscreen applications. For fullscreen applications, we should get the same performance.

A big plus that can get XWayland is being tearing free (which it isn't now). It should be possible (without impacting performance), but is quite complicated.

[UnkendTech]

mannerov の投稿を引用：

I think the title is a bit misleading.

If you look closely, the tests show much higher fps for Wayland when the fps of the benchmarks are already very high. It is because the Wayland compositor synchronizes to the screen refresh and won't do more than /refresh-rate/ copies per seconds.
So yes, fps are higher for high fps applications, and only the most recent frame, before the compositor draw the next frame for the screen, will be picked up and displayed.

However, for fps around or lower than 60 fps, X and the Wayland compositor would do about the same number of copies, and performance is similar.

People said than when XWayland would be entirely ready, its performance would be better than X, but it will be true only for non fullscreen applications. For fullscreen applications, we should get the same performance.

A big plus that can get XWayland is being tearing free (which it isn't now). It should be possible (without impacting performance), but is quite complicated.

i was looking at the ms and it look like to me Wayland will be way better for gaming

[mannerov]

The ms are lower because the gpu is less busy doing copies outside the application

[JamieLinux]

Doc my question as i havent really kept up on wayland.. Anytype of wayland equivelent of xforward eg for usefull things like remote desktop though flux?

[UnkendTech]

Jamie の投稿を引用：

Doc my question as i havent really kept up on wayland.. Anytype of wayland equivelent of xforward eg for usefull things like remote desktop though flux?

RealVNC, FreeRDP, are being worked on and, a few more. the network transparency in Xorg has been mostly dead any ways but yes there are a few on the way for Wayland

[zigfreed#1531]

Doc Holliday の投稿を引用：

Jamie の投稿を引用：

network transparency in Xorg has been mostly dead

It's completely dead. The Real Story Behind Wayland and X