next up previous contents
Next: nCUBE 6410 Up: Distributed Memory Algorithms Previous: Distributed Memory Algorithms

Volume Rendering on a Hypercube Architecture

This section will describe a parallel algorithm for message passing MIMD machines that was implemented on a nCUBE 6410 super-computer. The algorithm is based on dividing work based on the image space among a set of statically allocated clusters. The following sections describe the machine architecture and the algorithm in detail. The algorithm and the implementation described is due to Misericordia et al. [24].





Claudio Silva
Thu Apr 20 16:03:37 EDT 1995