Hey Stephan, Towards the end of my work on the Raspberry Pi, I had to adjust this area of the code to get banging performance. I can't remember the exact combination that was winning, but unified timers was horrid and I had to short circuit several other functions which effectively set a broken state by default.
On top of all of this, I saw weird partial updates on all the OMAP SGX based devices I tested against Qt 5. I have a beagle board video uploaded which shows the screen broken up into 4 discrete quadrants which are alternately updated. The EGLFS defaults/baseline really needs to be adjusted to be optimal out of the box. It requires to much experimentation to chance upon the magic combination at present, and I have dropped the ball on getting this adequately quantified and reported upstream. Cheers, Donald On Mon, Nov 19, 2012 at 9:43 PM, Stephan Kanthak <sty...@gmx.de> wrote: > Hi all, > > as promised some feedback on this matter... > > On 11/16/2012 07:37 AM, Stephan Kanthak wrote: >>> Are you doing "export QML_FORCE_THREADED_RENDERER=1"? Because afaik >>> without that you won't get the threaded renderer. Most issues I've seen >>> with animations have been due to not using the threaded renderer. > Just "export QML_FORCE_THREADED_RENDERER=1" subjectively works best for > me so far. I still do see tearing all over, so there's definitely no > waiting for VSYNC, either in Qt5/EGLFS or the TI driver, but animations > are smooth. > >>> You might also want to try the customcontext renderer from >>> ssh://codereview.qt-project.org:29418/playground/scenegraph.git >>> After building and doing make install there you should be able to use it >>> by doing "export QMLSCENE_DEVICE=customcontext" or running with >>> --device=customcontext > I tried this code, too, although I had some problems adding Animators > objects at the right spot in my QML files. It finally seems to work for > me once Animations + animated objects are both from Animators. However, > the result is not much better than with just "export > QML_FORCE_THREADED_RENDERER=1". Tearing is even more obvious as the > buffer swapping then seems to happen right in the middle of an animated > texture covering the lower 50% of the display. > >>> If it's a timing issue you might also try >>> QML_FIXED_ANIMATION_STEP=1 > After setting this, animations are a bit quicker, but smoothness also > wasn't better than with just "export QML_FORCE_THREADED_RENDERER=1". > > So, overall my remaining problem seems to be that buffers are not > swapped with VSYNC. Any idea wether this is due to Qt5/EGLFS or rather > the TI driver? > > Best, > Stephan > > _______________________________________________ > Interest mailing list > Interest@qt-project.org > http://lists.qt-project.org/mailman/listinfo/interest _______________________________________________ Interest mailing list Interest@qt-project.org http://lists.qt-project.org/mailman/listinfo/interest
