Parallel rendering of curvilinear volume data

Abstract Several volume rendering techniques have been proposed to visualize regular three-dimensional cartesian data sets. Often curvilinear grids that model a problem more accurately are used for analysis, but few techniques are available to render curvilinear data sets. This work presents a new approach to render cylindrical polar gridded volumes that are frequently used in astrophysical system models and computational fluid dynamics. Since the curvilinear grid that serves as input to the rendering system is the same as that used for analysis, it also provides the ability to directly link the rendering and simulation process. Parallel implementation of the algorithm in both SIMD and MIMD environments is also explored.