basic VBO support with vertex array fallback is now in as of r203. For those who don't know what a VBO is, its a way to load geometry into video memory so it doesn't have to cross the AGP/PCIx bus every frame. Speedup is on a scale of 200%-400% over vertex arrays, which are themselves a speedup of 200%-600% over display lists (which Soya uses).
... so yea, we're going to have to benchmark pysoy in a "full environment" and post the results to show, clearly, how crappy Soya is ;-) please test on your systems to ensure it works. So far I've tested on Rage128 (free), Radeon 9200 (free), and TNT2 (proprietary), all run slick as a whistle if you ignore the lack of time stabilization on the physics sim. next is editable geometry. _______________________________________________ PySoy-Dev mailing list [email protected] http://www.pysoy.org/mailman/listinfo/pysoy-dev
