Parallel rendering techniques for massively parallel visualization

As the resolution of simulation models increases, scientific visualization algorithms which take advantage of the large memory and parallelism of massively parallel processors (MPPs) are becoming increasingly important. For large applications, rendering on an MPP tends to be preferable to rendering on a graphics workstation, due to the MPP's abundant resources: memory, disk and numerous processors. The challenge is developing algorithms that can exploit these resources while minimizing overheads, typically communication costs. This paper describes recent efforts in parallel rendering for polygonal primitives as well as parallel volumetric techniques. It presents rendering algorithms, developed for MPPs, for polygonal, spherical and volumetric data. The polygon algorithm uses a data-parallel approach, whereas the sphere and volume rendering use a MIMD approach. Implementations for these algorithms are presented for a Thinking Machines Corp. CM-5 and a Cray Research Inc. T3D.