This project goal is to investigate applications of stream- based functional programming methodology to the implementation of multiprocessor based real-time systems, and specifically to signal processing applications. The work is based on the Sisal programming language. It will use Sisal's stream data types, static analysis methods, and automatic mapping to map parallel computations on processors in distributed computer systems. The Paradigm compiler will be redesigned and extended to perform static analysis and mapping of continuous stream processing computations. A related objective of this effort is to develop analytical methods for performance modeling of stream-based computations on multiprocessor systems. ***
