Hardware Subdivision and Tessellation of Catmull-Clark Surfaces

We present a table driven approach to performing Catmull-Clark subdivision in parallel utilizing one thread per output mesh vertex. We apply the procedure one or two times in order to isolate extraordinary vertices and simplify the input patch structures consumed by the hardware unit responsible for adaptive tessellation. From this simplified mesh, we evaluate the performance Stam’s direct evaluation procedure, a curvature continuous patching scheme, and a tangent plane continuous approximation suitable for displacement mapping.