SBIR-STTR Award

This proposed effort will accomplish advanced development, i.e., systematic application of tools and techniques, of technology for high performance computing. The first phase will provide the proof-of-concept of a new technology: High Performance Geometric Theory of Diffraction (HPGTD) for complex systems. This HPGTD technology will be developed and demonstrated using portions of an existing Rome Laboratory (RL/ERPT) conventional scattering code, GEMACS, for General Electromagnetic Model for the Analysis of Complex Systems. Phase II will complete the development of HPGTD, use it tp parallelize the rest of the GEMACS GTD software, and deliver and document the new code for production use. The resulting HPGTD software will be scalable and portable to massively parallel processors, but also regress to serial execution on a single processor for baseline validation.

The first phase developed proof-of-concept of the geometric theory of diffraction (GTD) for high performance computational electromagnetics. This phase II effort will demonstrate the GTD technology for complex systems using massively parallel processors and develop and demonstrate the other computational electromagnetics technologies for other domains. The multiple high performance computing technologies will be integrated and demonstrated using an operational software system, GEMACS, for General Electromagnetic Model for the Analysis of Complex Systems. The resulting HPC soft-ware will be benchmarked for scaling and porting to other massively parallel processors. Several high performance computing technologies will be used in this program: data parallelism, functional parallelism, and client-server over a wide area net. Wide are networking technologies will be used to provide distributed and remote operational capability; both the Internet to provide world-wide operation, and the New York State-based NYNET, to demonstrate asynchronous transfer mode (ATM) wide area distributed performance. Several areas for commercialization of these technologies have been successfully investigated to date. The phase II commercialization activities will focus on high performance computational electromagnetics technology for multi-disciplinary manufacturing, the Living Textbook project for education, and the common runtime support for high performance computing languages program.