Ubuntu Unity Proves Very Slow To KDE, GNOME, Xfce, LXDE

[Artemis3]

Actually, Xfce has a fully supported and configurable compositor, it is just not on by default. On 4.8 it can be toggled from the "Window Manager Tweaks" settings option under the "Compositor" tab. From there you can also tweak the compositor effects and also specifically set the compositor to "Display fullscreen overlay windows directly". Considering you took the time to tweak KDE for this benchmark, I would appreciate it if you would also do the same for Xfce in future tests.

No, the compositor is turned on by default in Xubuntu, which is why the benchmarks show lower FPS with XFCE than LXDE. Perhaps testing with both compositor on and off should clarify this. I rather use my desktop with the compositor off, the menu shadows and transparency of objects are not worth the tearing in videos and fewer FPS.

[Artemis3]

And so .. Was anyone saying KDE is heavy and slow and that XFCE is better for performance/gaming etc?

Do the benchmarks with a window instead of fullscreen, and the truth will come out. I often play rpg games windowed, as the need to browse documentation online comes very often. Also, the time KDE takes to boot is absurd, even worse than gnome3, which in itself is quite unbearable.

[Artemis3]

XFCE compositor can be turned off

If you turn off the compositor, XFCE will match the speeds of KDE suspend and LXDE.


Now i would like to see true desktop usage benchmarks. Time to boot, time to open the file browser, time to open a default app (libreoffice?), memory usage, power usage, fps on the desktop itself, etc.

[GreatEmerald]

More like someone else saying that the headlines are wrongly made. Which is a fair point.

[Ofen]

Hey Michael heres idea for your next article

“Look, someone else says my benchmarks are wrongly made”.

Source:
http://blog.martin-graesslin.com/blo...ce-benchmarks/

lol, shitstorm incoming

[Ramiliez]

Hey Michael heres idea for your next article

“Look, someone else says my benchmarks are wrongly made”.

Source:

http://blog.martin-graesslin.com/blog/2012/09/why-i-dont-like-game-rendering-performance-benchmarks/

[bug77]

an aswer from kde developer:

http://blog.martin-graesslin.com/blo...ce-benchmarks/

Sane remarks, but his points are valid for most benchmarks. Rarely will you see a piece of software benchmarked across several operating systems, CPU, etc. You can't realistically vary every single factor, so reviewers usually pick a few that (in their eyes) are the most important. Yet people will either generalize those results if they are inline with their views or dismiss the results as flawed if they contradict the same views.

[TomBoyTom]

Please test Pantheon Shell and Cinnamon

I like to see results with these two youngest WM

[gcala]

an aswer from kde developer:

http://blog.martin-graesslin.com/blog/2012/09/why-i-dont-like-game-rendering-performance-benchmarks/

[Rigaldo]

It does affect performance, but not with every driver. I'm running fglrx and there's no difference. With Open Source drivers and nVidia binary driver you should see a noticeable slow down with compositions being enabled. Some Phoronix benchmarks are so moronic, because they don't really benchmark driver speed (the best examples are Linux vs OS X and Intel hardware). I bet Kubuntu with suspended compositions turned on will simply kill crAppple OS.

Maybe in discrete cards the difference is just more negligible due to their higher "horse power", but from what I read in this thread previously I understand that composition puts a bit more hardware(gpu and cpu) overhead. On a discrete card(like mine) I wouldn't expect such a significant difference though, as the one shown here.

Or it is the driver failing when there's no composition so you don't get the gain(with composition turned off) instead of not getting the slow down(with composition turned on) actually.

Regarding the last one, I recall some benchmarks showing Ubuntu faster than OS X with NVIDIA hardware (if I recall correctly). And it was probably with compiz(and Unity? don't remeber ..).