Ubuntu 13.04 Desktop Gaming Performance Comparison

[gururise]

we need a universal call, like Windows has, for programs to be able to tell the WM to disable compositing when needed.

Best idea I've heard yet!! Now why doesn't something like this exist???

[stqn]

Best idea I've heard yet!! Now why doesn't something like this exist???

- it’s not required to make the games run
- I don’t want game developers to decide if my screen should be composited or not
- game developers don’t want to bother with this
- why ask game devs to code this in each and every game when it can (should) be solved once and for all in the WM?

[Luke_Wolf]

Agreed. Most Linux users are going to be using default settings.

To those who are complaining about these benchmarks: Do you really think it is realistic for the average user coming from Windows to know how to tweak numerous settings in each Desktop? I've used Gnome-Shell since its release, and still not sure how to "Unredirect fullscreen", or if there is even an option.

Again... Not Numerous Settings... one setting. Which I don't know about Gnome Shell (you probably have to use it's registry tool), but in KDE it's very easy to get to System Settings -> Desktop Effects -> advanced, and the bam you just check the "Suspend desktop effects for full screen windows" and you're done. Besides if they're doing gaming to the point where performance actually matters assuming they have a modern system (Llano/Sandybridge and forward) if they're using integrated graphics or if they've got at least a decent graphics card otherwise (it's insane to assume gaming period on systems that don't fit one of those two categories) you expect them to have at least some level of knowledge, not Joe Bloggs who isn't configuring anything.

[Hamish Wilson]

I've used Gnome-Shell since its release, and still not sure how to "Unredirect fullscreen", or if there is even an option.

Mutter already does this by default.

[Vim_User]

These benchmarks clearly show that performance of games can be influenced by compositing WMs.
So the real question that remains for me is: Why is Michael using Unity when it comes to benchmarks of different driver versions, when it is so obvious that there is a non-predictable influence on performance? Why not use Openbox for those benchmarks instead, which constantly delivers the same good performance?

[RussianNeuroMancer]

Best idea I've heard yet!! Now why doesn't something like this exist???

Exist:
https://projects.kde.org/projects/kd.../client.h#L262
https://projects.kde.org/projects/kd...ient.cpp#L2313

[blackout23]

Michael must have been pissed that the KWin dev mocked his benchmarks in his blog so he enabled desktop effects in fullscreen and did not unredirect fullscreen.

No everyone will be like ZOMG! KDE suxxx!