OpenGl implements depth (Zbuffer) visibility determinations.
Is there really some performance boost if I add BSP code
to my OpenGL code (used to visualize .OBJ models).
Let''s say I want to display a Quake Map, (or any kind of game
level). Isn''t it enough to let OpenGL do the rendering stuff,
do I have to use BSP or some algorithm to improve performance?
I have a TNT2-m64 card. Can a BSP tree make really a diferrence?
you need to handle visibility determination yourself. (lol, that thread is joke? isn''t it, i hope it isn''t) yeah imho BS-trees+VIS is best choice, cause your rasterizer needs to only take visibility stuffs from look-up-table and that''s quick, no calculations needed at all. And for faster processing, please cull stuffs that isn''t visible in vis still with cone of vision.