Polymorphous Computing Architectures (PCA)

Program Manager: Dr. William Harrod

Objectives:

Polymorphic is defined as having, taking, or passing through many different forms or stages (i.e. many + form). The emphasis of this program will be on polymorphous computing architectures and application development environments. The PCA program will establish the ability to span a broad dynamic application space by implementing a transparent reactive layer between the application program and the malleable micro-architecture elements. This polymorphic layer will enable software and hardware to be developed in a cooperative constraint sensitive environment instead of in a failure prone hardware first and software last paradigm. The resulting embedded computing systems will enable optimization across a broad range of applications and possess the ability to react to dynamic mission requirements.

The PCA program will implement a family of novel malleable micro-architecture processing elements to include compute cores, caches, memory structures, data paths, network interfaces, network fabrics with incremental instructions, OS, and network protocols. These elements will have the ability to reconfigure to match changing mission and scenario demands. To support the use of polymorphous computing systems, the program will create a model based software framework for reactive monitoring, optimization, modeling, resource negotiation and allocation, regeneration, and verification. A set of measurement metrics will be developed to support processing system design and optimization to include size, weight, energy, performance, and time (SWEPT). Specific PCA program goals are to allow post-silicon optimization through the incorporation of polymorphous concepts within commercial processing R&D and fabrication infrastructure; to develop an environment that provides resource allocation, negotiation, and monitoring; to implement verification and validation at multiple system levels; to develop testbeds and conduct proof-of-concept experiments; to facilitate technology transitions using strategic teaming; and to establish benchmark and standards groups creating community standards that enable broad application and commercial support of PCA program developments.

The mission payoff is possessing the ability to react to collaborative information centric strategies using a common highly optimized (based upon SWEPT) processing architecture regardless of the mission dynamics or sensor suites. The result will be the capability to provide optimized mission processing implementations (pre-mission, in-mission, and post-mission) without the necessity of custom development effort including associated time and cost.

Click Here for Figure One

PCA Architecture Implementation Environment

Click Here for Figure Two