A Dynamic Scheduling Algorithm Based on GaBP

摘要：

　　Based on Gaussian Belief Propagation (GaBP) algorithm for solving sparse symmetric linear equations, an iterative acceleration optimization method of G-aBP is studied.We explore the parallelism and load balancing features of this algorithm and present a multicore-based parallel GaBP algorithm with dynamic scheduling.The numerical results indicate that this algorithm can solve large scale sparse symmetric linear equations with high convergence and higher parallel efficiency.

关键词： parallelization GaBP algorithm dynamic scheduling sparse linear equations

作者: Hanyuan Zheng Anping Song Lei Xu Zhixiang Liu Minchao Wang Wu Zhang

作者单位: School of Computer Engineering and Science, Shanghai University,200072 Shanghai, China;School of Com School of Computer Engineering and Science, Shanghai University,200072 Shanghai, China

会议类型: 国际会议

会议名称: International Workshop on Data-Intensive Scientific Discovery and Applications 2013(2013数据密集型的科学发现与应用国际研讨会)

会议地点: 上海

会议语种:英文

页码: 183-189

在线出版日期: 2013-08-01（万方平台首次上网日期，不代表论文的发表时间）