Composition Bypass Support Lands To Speed Up Mir

[Weegee]

lets see how well DOTA 2 run's on it ohh wait nvm....

What's wrong with the DOTA 2 performance? It runs just fine on my Laptop (Mobility Radeon HD 5650) which uses the radeon drivers ...

[LinuxGamer]

The strange thing is... I actually prefer to run games without bypassing the composition.
Allows me to multi-task without funny stuff.

From my point of view, this is just a "fix" for benchmarking, and nothing to do with actual experience improvement.

Xmir will give no real "experience improvements" at all

In development software is in development.

Do out in only a few weeks and it has a Huge Hole in it too Remember VT Switching? and October 17 is coming a little fast dont you think? by that time we will have a fully ready Wayland Desktop for testing and Not some Xwayland patched up bull shit if the Wayland developers did some bull shit like this i be all over them but really the said part is we expect this from Canonical

What's wrong with the DOTA 2 performance? It runs just fine on my Laptop (Mobility Radeon HD 5650) which uses the radeon drivers ...

on Xmir? or Xorg fall back? i have yet to get this working on a Newer AMD GPU or Nivdia

[fossy]

no you will notice a huge difference when a game that play's at 30fps on xorg only plays at 23fps on Xmir Humans can tell a difference all the way up to 120fps+ this is a slap in the face to all of Ubuntu users
i

hmm - 6% of 30fps is a 2fps loss.

If you are playing a game at 28fps you need a better graphics card!

[Weegee]

on Xmir? or Xorg fall back? i have yet to get this working on a Newer AMD GPU or Nivdia

I'm running the unity-system-compositor, so on Xmir. It runs with around 20 fps, which is okay because I'm not playing on my Laptop anyway, I was just testing it.

[LinuxGamer]

hmm - 6% of 30fps is 2fps.

If you are playing a game at 28fps you need a better graphics card!

on a light weight game it will be 6% but on a Heavly Game that 6% can be 10% 20% or even 50% i want to be some real test and Btw how do we know what they really did to this to get them mark's you can tweak Xorg server to make it run a hell of a lot faster but it's going to be Unsafe and Unstable you know what i can gut the Xorg Server too (done it befor)

[fossy]

are we making figures up now? 50%

Michael says Canonical are using his software for benchmarking. Surely you are not saying his software is giving that amount of variation?!

[jrch2k8]

well XMir "optimizations" are not compatible with XWayland because XWayland is rootless while Xmir is a whole Xorg server with a glorified page-flipper + optional compositor[with this patch].

forgot to quote OP, see page 1

[fossy]

well XMir "optimizations" are not compatible with XWayland because XWayland is rootless while Xmir is a whole Xorg server with a glorified page-flipper + optional compositor[with this patch].

forgot to quote OP, see page 1

Yeah - I get the impression that TheOne is probably a 'buntu user.

Still waiting for a major distro such as Fedora to have Wayland stuff built in to benchmark against. Anyone have any dates when we should expect to have something alpha'ish to benchmark against?

[LinuxGamer]

are we making figures up now? 50%

Michael says Canonical are using his software for benchmarking. Surely you are not saying his software is giving that amount of variation?!

i'm not at all saying at all it's his software but i like to see him do the Marks and not them there is no telling what they did to the Xorg Server to get this marks and look at the marks they ran "light weight's"

[fossy]

i'm not at all saying at all it's his software but i like to see him do the Marks and not them there is no telling what they did to the Xorg Server to get this marks and look at the marks they ran "light weight's"

knowing michael's propensity for this stuff ... he is probably benchmarking this right now :P