I'm working on PhoenixGL's optimizing, batching engine. This is a simple test of the concept. There is no caching here, and no sort of tree in place to split and optimize batches. This is just a test of one batch of sprites at a single depth. Each 'sprite' transforms and passes its vertex (position, color, tcoords) to a batch renderer which just pushes them all with glDrawArrays. This produces a 10x increase in framerate over immediate rendering. This batch shares three properties in common: same depth, same texture, and same geometric primitive (GL_QUAD). These three criteria will be the leaves of the tree that will eventually become a full-blown batching renderer.
Link to this comment:
All Comments (0)