Hello Aleksander, Aleksander Stroiński wrote: > Hello, > > I am doing some tests with sort-last rendering. > For my test scene (5 milion polygons) a got results: > > number of nodes | fps | speedup > > 1 | 0,349180722 | 1 > 2 | 0,973361065 | 2,787556708 > 3 | 2,335813148 | 6,689410379 > 4 | 2,999614256 | 8,590434864 > 5 | 3,092690264 | 8,856990284 > 6 | 3,132578463 | 8,971223964 > 7 | 3,223628831 | 9,231978243 > 8 | 3,234177919 | 9,262189215 > 9 | 3,248149293 | 9,302201085 > 10 | 3,325776935 | 9,524514739 > > it is easy to see that speedup is SUPER LINEAR!
nice :) > how it is possible? > > i heard something about CACHE EFFECT - more nodes --> more memory > > My program has 1,5GB memory, but machine has 2,5GB RAM, so it isn't > cache effect... hm, one possibility is the reduced traversal overhead on each node (only a part of the scene graph has to be traversed on each node), also less data (and draw commands) has to be transfered to the gpu, etc. But really this is just guesswork and I'm not sure if there has been an analysis of the bottleneck(s) in the SortLast implementation - maybe in Markus Roth's PhD thesis (http://elib.tu-darmstadt.de/diss/000672/ - sorry it's in German) ? Hope it helps, Carsten ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ Opensg-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/opensg-users
