Realization of Replicated Streaming Applications on Multi-Clocked FPGAs

Streaming applications are used in many fields including telecommunications, multimedia and the Internet. Due to their inherent parallelism and high performance requirements, they are naturally implemented in hardware or multi-core systems. In this thesis, the streaming applications which are described in StreamIt languages are realized directly in FPGA based hardware. However, it is not an efficient implementation if the applications simply map to hardware. Hence, we adopt a replication algorithm and a multi-clock domain mechanism to increase the throughput of the applications. The results of experiments show that these two algorithms can optimize the throughput subject to area and clock frequency constraints. Furthermore, a trade-off is explored in these two methods