Development of High Performance Computing (HPC) Technology for a Distributed Modeling and Simulation (M&S) Hardware Infrastructure

This research effort will develop a heterogeneous, high-performance computing platform for distributed modeling and simulation applications capable of providing flexible configurations of CPU, GPU and FPGA resources that use a simple, common and open programming model (OpenCL). The use of OpenCL enables functionally equivalent algorithms to be deployed across CPU, GPU and FPGA resources. Deployment decisions can be postponed until run-time allowing optimization of the deployment for the given scenario (e.g. high performance, portability, low-power, etc.). This HPC platform is capable of supporting large enterprise-scale simulations, hardware-in-the-loop simulations, as well as portable applications. The platform is capable of leveraging various storage technology including large RAID configurations and Solid-State Drive (SSD) solutions. The overall solution will increase computational performance, while simultaneously reducing the power and space footprint. This research has immediate direct impact to current military needs, but also provides a solution to many commercial market needs including specific modeling and simulation applications (e.g. financial, genetic, weather) and as an enabling platform for High Performance Computing (HPC)-as-a-service. Approved for Public Release 14-MDA-8047 (14 Nov 14)