[Xcb] Question about xcb performance and threading support

Benoît Minisini gambas at users.sourceforge.net
Tue Feb 17 15:05:28 PST 2009


> HI Benoît,
>
> > I have currently using valgrind / callgrind & kcachegrind with the Gambas
> > IDE to try to understand why QT4.4 is slower than QT3.
>
> QT4 is now double-buffered, furthermore it uses antialiasing for a lot
> of its widgets and has a new painting engine which uses XRender
> trapezoids. Unfourtunatly the trapezoids are not hw accalerated, so
> they are slow compared to what QT3 did.
>
> I am not 100% sure, but when running your benchmarks pay attention to
> the time spent in Xorg in relation to the QT client.
>
> - Clemens

Sorry, I made a mistake in sending some previous answer to this post, so you 
didn't see it.

The xcb problem is the same with QT3 and QT4. QT4 just does more X requests I 
think. And with valgrind I don't see the time spent in Xorg, only the time 
spent by my process and its shared libraries.

In another answer, I posted a screenshot of kcachegrind with the relative time 
used by all xcb related function. Apparently the problem is in xcb-xlib that 
takes a lot of time in locking/unlocking the X11 display.

Regards,

-- 
Benoît


More information about the Xcb mailing list