- fvi_ray_boundingbox() is faster and should be less prone to compiler/optimization/platform specific issues - clean up or obj_check_all_collisions() which is a good bit faster now