An approach for multi-task and multi-application mapping onto NoC-based MPSoC

The application mapping on a multiprocessor platform can influence in the system performance and in the energy consumption of an embedded systems. In this context, this paper presents a meta-heuristic based approach for energy aware mapping an application into a NoC-based multiprocessor platform. The proposed approach was able to find a processor mapping that reduces the energy consumption of about 66%, in average, for the a multi-task application or a set of applications, in comparison with the worst case mapping.