Implementation of the Joint Analysis System on TOW to Enhance DoD Analysis Performance

• Published in: 2010 DoD High Performance Computing Modernization Program Users Group Conference pp. 327 - 332
• Main Authors: Sweetser, Al, Barnes, Boots, Melim, Peter, Bouwens, Christina, Pratt, David
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.06.2010
• Subjects: Analysis; Analytical models; Computational modeling; Computer architecture; Constructive Simulations; Forces Modeling; Hardware; High Performance Computing; Joint Analysis System; Joints; Linux; Simulation; US Department of Defense
• ISBN: 9781612849867; 1612849865
• DOI: 10.1109/HPCMP-UGC.2010.42

The Joint Analysis System (JAS) is a multi-sided, joint campaign-level simulation used by the Office of the Secretary of Defense/Director of Cost Assessment and Program Evaluation (OSD/CAPE) Simulation and Analysis Center (SAC) for theater-level strategic analysis to support and assist Department of Defense (DoD) decision-making. The need for a higher execution speed for JAS (400 times real-time) combined with a desire for the shortest possible time-to-solution has led to the application of high performance computers (HPCs) to support the JAS program goals. Previous work with JAS [previously called the Joint War fighting System (JWARS)] explored the viability of parallel execution of computationally-intensive simulation processes to shorten execution times. Further work was performed to port the "back-end" simulation code to the Linux® (Linus Torvalds) environment to take advantage of Linux-based HPC assets. Capabilities developed as part of the parallel processing effort are part of the JAS technical library, and the Linux porting is now part of baseline code. The US Army Research Laboratory (ARL) DoD Supercomputing Resource Center (DSRC) has recently added a 6,656 core SGI Altix® ICE 8200 system (called TOW) to its toolset. This new system offers the SAC an opportunity to further utilize the processing capabilities offered by HPCs. The focus of this paper is to discuss the challenges and lessons learned in adapting JAS for use on TOW.