Interactive Volume Rendering on Clusters of Shared-Memory Multiprocessors

Proceedings of Parallel Computational Fluid Dynamics '95, Pasadena, CA.

Authors:

Michael E. Palmer, California Institute of Technology

Stephen Taylor, California Institute of Technology

Brian Totty, Silicon Graphics Computer Systems

Abstract:

We describe methods for parallel partitioning and load balancing of volume rendering on a cluster of shared-memory multiprocessors; good load balance and parallel speedup are demonstrated. We describe a method of dynamic resolution reduction and image refinement which typically allows a four-fold performance increase.

Paper (not including Color plate 1)

Please save the following link as a file (shift-click in Netscape) to download the paper in compressed postscript (.ps.gz) format:

pcfd95.ps.gz (123KB)

Color plate 1

Since this color plate is large (17MB) in uncompressed form, it is not included in the postscript file above. You can either click here to view as HTML, which you may choose to print from Netscape, or alternatively shift-click here to download in compressed postscript (.ps.gz) format. (2.8MB)