In this paper we present a scheme for reducing the amount of synchronization overhead needed between components, after HW/SW partition-. ing, to preserve the original control flow of the specification. Since tra c between components is expensive, our scheme can significantly enhance the performance of the system implementation. Our'optimization technique dynamically groups the tasks in the specification such that synchronization for different tasks can be shared. The grouping depends on the partitioning decision, and hence, is performed during the generation of the partitioned model. We apply our grouping algorithm for various partitions on system level models of industry standard designs, The experimental results show significant reduction in synchronization overhead compared to the unoptimiaed model. 0-7803-8736-8/05/$20.00 02005 IEEE.

Message-passing mechanism is commonly used to preserve data coherency in distributed systems. This paper presents an algorithm for insertion of minimal message-passing in system-level design to guarantee data coherency. The target architecture is a multi-component heterogeneous system, where some components have local memory (or they are memory components by themselves. The algorithm enables automatic insertion of message-passing during systemlevel design to relieve designers from tedious and error-prone manual work. The optimal solution given by the algorithm also ensures the quality of automatic insertion. Experiments show that the automatic approach achieves a productivity gain of 200X over manual refinement.