Research on Parallel HW/SW Partitioning Based on Hybrid PSO Algorithm

Authors:
Yue Wu;Hao Zhang;Hongbin Yang
Affiliations:
School of Computer Engineering and Science, Shanghai University, Shanghai, China 200072;School of Computer Engineering and Science, Shanghai University, Shanghai, China 200072;School of Computer Engineering and Science, Shanghai University, Shanghai, China 200072
Venue:
ICA3PP '09 Proceedings of the 9th International Conference on Algorithms and Architectures for Parallel Processing
Year:
2009

Citing 4
Cited 0

A connectionist machine for genetic hillclimbing

A connectionist machine for genetic hillclimbing
An overview of message passing environments

Parallel Computing - Special issue: message passing interfaces
TGFF: task graphs for free

Proceedings of the 6th international workshop on Hardware/software codesign
Multi-objective optimization of interconnect geometry

IEEE Transactions on Very Large Scale Integration (VLSI) Systems - Special section on system-level interconnect prediction (SLIP)

Quantified Score

Hi-index	0.00

Visualization

Abstract

Generally, CHardware/Software (HW/SW) partitioning can be approximately resolved through some kinds of optimal algorithms. Based on both characteristics of HW/SW partitioning and Particle Swarm Optimization (PSO) algorithm, a novel parallel HW/SW partitioning method is proposed in this paper. A model of parallel HW/SW partitioning on the basis of PSO algorithm is established after analyzing the particularity of HW/SW partitioning. A hybrid strategy of PSO and Tabu Search (TS) is proposed in this paper, which uses the intrinsic parallelism of PSO and the memory function of TS to speed up and improve the performance of PSO. To settle the problem of premature convergence, the reproduction and crossover operation of genetic algorithm (GA) is also introduced into procedure of PSO. Experimental results indicate that the parallel PSO algorithm can efficiently reduce the running time even for large task graphs.