Partición hardware software de un codificador JPEG utilizando escalador de colinas estocástico

Abstract

Hardware/software partitioning is a key task for embedded system co-design. The goal of this task is to decide which components of an application will be executed in a general purpose processor (software) and which ones on a specific hardware. To support this decision a design space exploration is executed, by the evaluation of several solutions to establish the best trade-off reached. To accomplish this task, metaheuristics algorithms are used by the most proposals; highlighting Genetic Algorithms and Simulated Annealing. Many times this decision is not taken by a comparative study over several algorithms. In this article the application of Stochastic Hill Climbing and Restart Stochastic Hill Climbing for solving the hardware/software partitioning problem is presented. A case study of JPEG encoder is presented. The results show that comparable solutions are reached by those algorithms.

Keywords

• hardware/software partitioning
• JPEG encoder
• restart stochastic hill climbing
• stochastic hill climbing