Parallel volume rendering

During the last years several methods have been developed to visualize three dimensional volume data on multiprocessor computer systems. Most of these algorithms were strongly adapted to the underlying target architecture, and often did not take account of requirements such as scalability, portability, or memory minimization. The following paper outlines a new approach to rendering large scale volume data on distributed memory parallel computers. Within this framework, the only limiting factor concerning the resolution of the processed data is the node local memory. Furthermore, porting to other multiprocessor systems is straight forward, and scalability with increasing numbers of processors is achieved.