Title: Two-steps into energy consumption optimisation due to the mapping of multimedia application to network on chip architecture
Authors: Djalila Belkebir; Fateh Boutekkouk
Addresses: Department of Mathematics and Computer Science, Research Laboratory on Computer Sciences, Complex System (RELA(CS)2), Oum El-Bouaghi University, P.O. Box 358, 04000 Oum El-Bouaghi, Algeria ' Department of Mathematics and Computer Science, Research Laboratory on Computer Sciences, Complex System (RELA(CS)2), Oum El-Bouaghi University, P.O. Box 358, 04000 Oum El-Bouaghi, Algeria
Abstract: Energy-efficiency is becoming one of the most critical issues in embedded system design that target the multimedia application caused by the increasing number of intellectual property (IP) cores in network on chip (NoC). This paper addresses the optimisation of NoC performances in term of power and latency, we proposed an optimisation technique that is an hybrid based-scheduling algorithm that evolve the cellular automata (CA) with genetic algorithm to solve mapping and scheduling problems where we mixed between GA easiest implementation and achievability of near global optimum solutions with CA simplicity and rigorous mathematical model to achieve our goal. In our algorithm, each transition rule represents a chromosome allowing an automatically programming of the transition rules of the evolutionary cellular automata. Also we have presented new dynamic voltage and frequency scaling technique that automatically detect the voltage change point in order to save power while respecting the timing constraints of the soft real-time multimedia application.
Keywords: NoC; network on chip; cellular automata; genetic algorithms; mapping; scheduling; energy consumption optimisation; multimedia applications; energy efficiency; embedded system design; embedded systems; latency; voltage change point.
International Journal of Intelligent Systems Technologies and Applications, 2016 Vol.15 No.4, pp.353 - 378
Received: 22 Oct 2015
Accepted: 11 Apr 2016
Published online: 02 Nov 2016 *