Computing 2D Constrained Delaunay Triangulation Using Graphics Hardware

This paper presents a novel approach, termed GPU-CDT, to compute the constrained Delaunay triangulation (CDT) for a planar straight line graph (PSLG), consisting of points and edges, using the graphics processing unit (GPU). Although there are many algorithms for constructing the 2D CDT using the CPU, there has been no known prior approach using the parallel computing power of the GPU efficiently. For the special case of the CDT problem with PSLGs consisting of just points, which is the normal Delaunay triangulation problem, a hybrid approach has recently been proposed that uses the GPU together with the CPU to partially speed up the computation. Our GPU-CDT works for such special case too, but the whole computation is fully accelerated by the GPU. Our implementation using the CUDA programming model on nVidia GPUs is numerically robust and runs several times faster than any existing CPU algorithms as well as the prior GPU-CPU hybrid approach. This result is reflected in our experiment with both randomly generated PSLGs and real world GIS data, with millions of points and edges. CR Categories: I.3.5 [Computer Graphics]: Computational Geometry and Object Modeling—[Geometric algorithms]; I.3.1 [Computer Graphics]: Hardware Architecture—[Graphics processors]