Tuning Methodologies for Parameterized Systems Design

In this paper we present some new solutions for design space exploration of parameterized systems. The approaches use multi-objective optimisation techniques based on the concept of Pareto-optimality to determine the power/performance trade-off front for a highly parametric system-on-a-chip for digital camera applications. The approaches used are purely heuristic, genetic algorithm-based and a combination of the two. The results obtained demonstrate the effectiveness of the approaches in terms of both validity and efficiency, measured as the number of simulations run.