Elyas Khajekarimi - College of Engineering, University of Tehran
Mahmoud Reza Hashemi

Existing research confirms that proper mapping of applications and their corresponding tasks has a significant effect on both the performance and the energy consumption of NoC-based MPSoC platforms. The objective of this paper is to obtain an optimal mapping and scheduling of an application represented by its task graph, on a target MPSoC architecture such that the energy consumption is minimized while performance constraints such as application deadline are satisfied. In this work, both processing and communication energies have been considered. The optimal mapping solution is determined using an ILP approach. Experimental results indicate that the proposed scheme has been able to save energy by 7% in comparison to the ILP method with the aim of minimizing processing energy and to reduce energy by 17% in comparison to the ILP method that try to minimize communication energy. Energy reduction is more than 20% when compared to a mapping method that uses evolutionary algorithms such as the Genetic Algorithm

Task mapping, Integer linear programming, heterogeneous MPSoC