The SDK says that pn triangles are actually emulated on the RadeON. What I want to know is if the emulation is fast enought to be used in a realtime application, or is it faster to do my own subdivisions and therefore send more vertices to the card ?
Jason ? Anyone ? Thanks !
> What I want to know is if the emulation
> is fast enought to be used in a realtime
> application, or is it faster to do my own
> subdivisions and therefore send more
> vertices to the card ?
The emulation of PN Triangles on RADEON is not intended to be fast enough for real-time. You should pre-tessellate your models for real-time performance. A MAX exporter (with source) is forthcoming.
Obviously, this pre-tessellation is just for prototyping. Once there is a hardware implementation of PN Triangles, you will send your low poly meshes to it.
There is a simple PN Triangles sample available now.
[This message has been edited by JasonM [ATI] (edited 10-24-2001).]
Does the Radeon 8500 provide hardware support for pn_triangles?