Efficient run-time task allocation in reconfigurable multiprocessor System-on-Chip with Network-on-Chip

摘要

Due to the advancement of VLSI (Very Large Scale Integrated Circuits) technologies, we can put more cores on a chip, resulting in the emergence of a multicore embedded system. This also brings great challenges to the traditional parallel processing as to how we can improve the performance of the system with increased number of cores. In this paper, we meet the new challenges using a novel approach. Specifically, we propose a SOPC (System on a Programmable Chip) design based on multicore embedded system. Under our proposed scheme, in addition to conventional processor cores, we introduce dynamically reconfigurable accelerator cores to boost the performance of the system. We have built the prototype of the system using FPGAs (Field-Programmable Gate Arrays). Simulation results demonstrate significant system efficiency of the proposed system in terms of computation and power consumption. Our approach is to develop a highly flexible and scalable network design that easily accommodates the various needs. This paper presents the design of our NoC (Network on Chip) which is a part of the platform that we are developing for a reconfigurable system. The major drawback of SOPC based systems lies in the routing of the various on-chip cores. Since it is technically difficult to integrate more than one core on a single chip, we come across several routing problems which lead to inefficient functioning. Thus we implement a NoC based routing algorithm, power Aware topology algorithm which considerably improve accessing speed and enhance the system efficiency, with nearly 85% of conservation of energy and efficient run-time task allocation of the system.