This software innovation speeds up the computation time of the Satellite Orbit Analysis Program (SOAP) tool by parallelizing the code using the message- passing interface (MPI) library. The speed increases almost linearly with the number of processors, allowing the perstudy duration of the visualization and analysis of space missions to take place in hours, rather than overnight. The software can conduct a parametric study involving millions of design vectors in a few hours of computational time by distributing the design scenarios among multiple processors. This allows SOAP to run in a parallel mode on JPL’s new, high-performance computer cluster, which has 1024 Intel Xeon processors.
Parametric Study and Contours in SOAP were the first targets of parallelization. The two functions compute one or more variables over a given time period and become quickly computationally intensive as report duration, time resolution, and variable complexity increase. The spatial and variable domains are sliced and distributed to each processor proportionally to its processing power. The result of the computation over the slice is collected at the end of the computation, and a single processor handles the file-writing task. Each space and variable domain contribution is completely independent so that message passing is needed only at the beginning and end of the computation. The minimal message passing ensures the high speed of the proposed parallel implementation.
Trade analyses can be performed and optimization problems can be solved for space mission design with respect to relevant figures of merit. The mission merits/objectives/requirements are calculated with the set of design variables. The tradeoff space given by the design problem is visualized. SOAP is used for DAWN mission design, Cassini mission analyses, Team X mission concept studies, and DOD and Air Force projects.