Personnel
Computational Engineering and Computer Sciences Team earns high marks for solving complex engineering problems
By Dale Shires, Computer Scientist
The goal of computational engineering is to provide insights and solutions for numerous classes of engineering problems. The complexity and size of the models used to solve these problems bring into focus the need for large, scalable computing platforms. However, code efficiencies can easily degrade when scaling to such sizes if key aspects of computer science for high performance computing (HPC) are not taken into account.
 |
| From left to right: Brian Henz, Dan Pressel, Dave DeMay, Dale Shires, Dixie Hisley and Csaba Zoltani. |
The Computational Engineering and Computer Sciences Team, part of the Computational Sciences and Engineering Branch of the High Performance Computing Division, brings together the right mix of interdisciplinary talent to effectively utilize these HPC resources. The team consists of Brian Henz, Mechanical Engineer; Dixie Hisley, General Engineer; Csaba Zoltani, Research Physicist; Dan Pressel, Computer Scientist; Dave DeMay, Contractor; Sha’Kia Boggan, Science and Technology Academic Recognition System student; and me, a Computer Scientist.
Henz and I are working on customer-focused research dealing with virtual manufacturing, prototyping, and testing. We are completing the final stages of development of the scalable PhoenixFlow code (developed under the auspices of the DoD’s High Performance Computing Modernization Program Office (HPCMPO) Common HPC Software Support Initiative [CHSSI]).
PhoenixFlow is a multidisciplinary application encompassing fluid flow, heat transfer, and resin cure kinetics as they occur in composite material fabrication processes. Extensions include coupling the results of PhoenixFlow with a new code we developed for multi-scale residual thermal stress analysis (MSStress). These results will be used by codes such as DYNA3D or p-DINOSAURUS for dynamic loading analysis.
“Computing hardware is constantly changing, challenging us to develop portable software while maintaining efficiency,” said Henz. “But it is so exciting to be able to utilize the computing power available to us to solve larger and more complex problems every day.”
Zoltani uses parallel computing assets at the ARL MSRC to perform calculations on threat agents and electrophysiological interactions. He collaborates extensively with researchers at the U.S. Army Medical Research Institute of Chemical Defense (USAMRICD) and is using scalable HPCs to impact this important area. He has also co-authored a chapter about computer simulations of threat agent effects on cardiac toxicity in the book, "Pharmocological Perspectives of Some Toxic Chemicals and Their Antidotes."
Hisley recently completed her doctorate's degree in Computer Science at the University of Delaware. Her thesis dealt with analysis and optimization of OpenMP parallel codes with programmer-controlled memory consistency models. This research provided new insights into how classical compiler analysis and optimization passes can be used in a parallel programming framework. Hisley is also contributing to profiling and optimization efforts for several codes running at the ARL MSRC.
Pressel's primarily responsibility is optimizing the performance of scientific codes on parallel architectures found at the ARL MSRC and other sites.
He is working on a new approach to estimate HPC performance when running computationally intensive programs. The technique, known as ENVELOPE, can be useful as a guide to select system configurations for procurement and help users select the best systems for their applications. The preliminary results have been promising, and Pressel continues to work with associated researchers to develop approaches to attack this difficult problem.
DeMay has been developing a general-purpose web publishing framework for HPC projects. Customer focus, feedback, and interaction have been important components to the success of parallel code development in the CISD. He is developing tools that incorporate internet technologies to allow researchers to increase customer interaction and facilitate all managing software development projects by merging web technologies, databases, and scripting languages as well.
Boggan is completing her senior year at South Carolina State University. During breaks, Sha’Kia is at the ARL MSRC investigating the use of object-oriented software design for HPC. Upon completion of her undergraduate degree, she will be pursuing her master’s degree in computational science.
The team continues to look forward to the new challenges and opportunities that emerging technologies and HPC resources bring. The rewards can be substantial.