Informatics Publishing Limited

M. Jagadeeswari ¹, M. C. Bhuvaneswari ²

Affiliations
1 Sri Ramakrishna Engineering College, Coimbatore, IN
2 P.S.G College of Technology, Coimbatore, IN

Abstract

Hardware software partitioning deals with the decision to partition a system description to be more suited to be implemented in special purpose hardware or software running on a standard processor. This is the key task of hardware software co-design, as the decision made at the early stage of the design process impact directly on the performance and cost of the system. This paper presents a novel application of Binary Particle Swarm optimization (BPSO) algorithm for hardware software partitioning. The algorithm operates on functional blocks for designs represented as Directed Acyclic Graph (DAG) with the objective to obtain a Hardware or Software implementation that meets performance requirements with a reduced design cost. Test problems are constructed randomly and the optimal solutions obtained from BPSO algorithm are compared with the optimal solutions obtained from traditional genetic algorithm. Experimental results show that BPSO is capable of finding optimal solutions very fast.

Keywords

Embedded Systems, Particle Swarm Optimization, Genetic Algorithms, Hardware Software Partitioning.

Full Text

     

M. Jagadeeswari ¹, M. C. Bhuvaneswari ²

Affiliations
1 Sri Ramakrishna Engineering College, Coimbatore, IN
2 P.S.G College of technology, Coimbatore, IN

Abstract

Hardware software partitioning deals with the decision to partition a system description to be more suited to be implemented in special purpose hardware or software running on a standard processor. This is the key task of hardware software co-design, as the decision made at the early stage of the design process impact directly on the performance and cost of the system. This paper presents a novel application of Binary Particle Swarm optimization (BPSO) algorithm for hardware software partitioning. The algorithm operates on functional blocks for designs represented as Directed Acyclic Graph (DAG) with the objective to obtain a Hardware or Software implementation that meets performance requirements with a reduced design cost. Test problems are constructed randomly and the optimal solutions obtained from BPSO algorithm are compared with the optimal solutions obtained from traditional genetic algorithm. Experimental results show that BPSO is capable of finding optimal solutions very fast.

Keywords

Embedded Systems, Particle Swarm Optimization, Genetic Algorithms, Hardware Software Partitioning.

Full Text

     

G. Subashini ¹, M. C. Bhuvaneswari ²

Affiliations
1 Department of Information Technology, P.S.G. College of Technology, IN
2 Department of Electrical and Electronics Engineering, P.S.G College of Technology, IN

Abstract

To meet the increasing computational demands, geographically distributed resources need to be logically coupled to make them work as a unified resource. In analyzing the performance of such distributed heterogeneous computing systems scheduling a set of tasks to the available set of resources for execution is highly important. Task scheduling being an NP-complete problem, use of metaheuristics is more appropriate in obtaining optimal solutions. Schedules thus obtained can be evaluated using several criteria that may conflict with one another which require multi objective problem formulation. This paper investigates the application of an elitist Nondominated Sorting Genetic Algorithm (NSGA-II), to efficiently schedule a set of independent tasks in a heterogeneous distributed computing system. The objectives considered in this paper include minimizing makespan and average flowtime simultaneously. The implementation of NSGA-II algorithm and Weighted-Sum Genetic Algorithm (WSGA) has been tested on benchmark instances for distributed heterogeneous systems. As NSGA-II generates a set of Pareto optimal solutions, to verify the effectiveness of NSGA-II over WSGA a fuzzy based membership value assignment method is employed to choose the best compromise solution from the obtained Pareto solution set.

Keywords

Heterogeneous Computing, Task Scheduling, Multi-Objective, NSGA-II, Pareto-Optimal.

Full Text