The Rationale of Distributed Rendering Using OpenGL in a Parallel CFD Environment

摘要

The advent of parallel and distributed processing has led researchers to seek new and innovative approaches to visualization of large memory, time dependent problems such as those encountered in computational fluid dynamic (CFD) applications. The current approach uses a concept whereby graphic calls can be issued at remote sites in a distributed parallel environment and appear within a single window on a workstation. This approach eliminates the need for explicit data migration to a centralized site for purposes of visualization. This paper describes a prototype implementation of a distributed CFD simulation system using the concept of distributed rendering. The rationale of this design is compared to an approach in use at NASA Ames which encapsulates the graphics entities into extracts which are then transferred to a remote client process and rendered locally.