System synthesis of synchronous multimedia applications

Abstract

Modern system design is being increasingly driven by applications such as multimedia and wireless sensing and communications, which have intrinsic quality of service (QoS) requirements, such as throughput, error-rate, and resolution. One of the most crucial QoS guarantees that the system has to provide is the timing constraint among the interacting media (synchronization) and within each media (latency). We have developed the first framework for system design with timing QoS guarantees. In particular, we address how to design system-on-chip with minimum silicon area to meet both latency and synchronization constraints. The proposed design methodology consists of two phases: hardware configuration selection and on-chip memory/storage minimization. In the first phase, we use silicon area and system performance as criteria to identify all the competitive hardware configurations (i.e., Pareto points) that facilitate the needs of synchronous applications. In the second phase, we determine the minimum on-chip memory requirement to meet the timing constraints for each Pareto point. An overall system evaluation is conducted to select the best system configuration. We have developed optimal algorithms that schedule a priori specified applications to meet their synchronization requirements with the minimum size of memory. We have also implemented on-line heuristics for real-time applications. The effectiveness of our algorithms has been demonstrated on a set of simulated MPEG streams from popular movies.