A parallel hardware/software partitioning method based on conformity particle-swarm optimization with harmony search

Abstract

Hardware/software (HW/SW) partitioning is a key step in HW/SW codesign. With the increasing design complexity of embedded systems, HW/SW partitioning has become a challenging optimization problem. A parallel HW/SW partitioning method based on Conformity Particle-Swarm Optimization with Harmony Search (CPSO-HS) is presented in this paper. Firstly, the particles act as psychological conformists and tend to move towards a security point, with many particles and a lower possibility of being attacked by a predator. By simulating the conformist mentality in CPSO-HS, the searching population can remain varied and the algorithm avoids local optima. Secondly, to improve the initialization strategy, the Harmony Search (HS) is integrated to search for better positions, with which the global best position is updated. Hence, the searching precision and solution quality can be enhanced. The searching diversification and intensification in CPSO-HS can be improved through the two methods. Thirdly, since the time to compute the HW/SW communication cost is the most time-consuming process for the HW/SW partitioning problem, we adopt a multi-core parallel technique to accelerate the computing. Thus, the CPSO-HS runtime for large-scale HW/SW problems can be reduced efficiently in an ordinary PC platform. Finally, a number of experiments on benchmarks from state-of-the-art publications demonstrate that the proposed approach can achieve higher performance than other previous methods.