SHEMUS - SYNTHESIS OF HETEROGENEOUS MULTIPROCESSOR SYSTEMS

Current VLSI technology enables the implementation of a complicated sy stem on a single chip at a low cost. Thus it has become cost effective to design special-purpose multiprocessor architectures for computatio nally intensive applications in signal processing, control of power sy stems and robotics. This paper presents a strategy, designated as SHEM US, for the synthesis of application-specific heterogeneous multiproce ssor systems to meet the various cost and performance constraints. SHE MUS combines a known fast heuristic with a standard genetic algorithm to search a large design space efficiently and effectively. The effect iveness of our technique is demonstrated by comparing it with some exi sting systems. The proposed strategy provides considerable improvement s in the cpu times for reasonable sized problems over previous work.