Artificial Intelligence Enhanced Parallel Computing Environmental for Real-Time Information Processing

The information processing tasks associated with defense application, such as Automated Target Recognition (ATR), have very diverse computational requirements that result in different needs for computing systems capabilities. A "heterogeneous" parallel computing environment provides a variety of architectural capabilities, orchestrated to perform an application whose tasks have such diverse execution requirements. One form of a heterogeneous computing system is a "mixed-mode" or "reconfigurable" system, where a single parallel machine can operate in different modes of parallelism. Another form is a "mixed-machine" system, where a suite of different high performance machines are connected together by high-speed links. A key issue that must be addressed in using heterogeneous parallel computing architectures for real-time information processing is the design of a high-level operating system for matching the processing tasks to the appropriate machine in a mixed-machine parallel system and to the appropriate architecture configuration in a mixed-mode/reconfigurable parallel system. This effort seeks to establish the feasibility of a new artificial intelligence (AI) based technique for implementing a high-level operating system for heterogeneous parallel computers. The proposed intelligent operating system consists of architecture independent, generalized routines that are useful in all parallel computing environments and architecture specific specialization for a given architecture in the form of rules in the rule-base of a knowledge-based system.