期刊文献+

求解大规模稀疏线性代数方程组序列的自适应AMG预条件策略

An adaptive AMG preconditioning strategy for solving large-scale sparse linear systems
收藏 分享 导出
摘要 时间相关偏微分方程隐式离散后,通常需要求解一个稀疏线性代数方程组序列.利用序列中相邻方程组性质的差异性与相似性,自适应地选取预条件子,提升方程组序列的并行求解效率,从而缩短总体求解时间,是一个值得研究的问题.本文针对科学与工程计算中广泛使用的代数多重网格(AMG)预条件子,设计了方程组序列相关的自适应预条件策略.通过惯性约束聚变(ICF)的辐射流体力学数值模拟典型应用,验证了该策略的有效性.测试结果表明,在某高性能计算机的3125个CPU核上,自适应预条件策略可将并行效率从47%提升到61%,将模拟总时间从19.7 h降为14.5 h. A series of sparse linear systems must be solved in applications that are based on the implicit solution of time-dependent partial differential equations(PDEs).Preconditioned iterative methods are usually employed to solve such sparse linear systems.AMG is one of the most popular preconditioners in real applications.However,it results in poor parallel scalability,owing to its setup phase.In this paper,by utilizing the differences and similarities in property among the systems in series,an adaptive AMG preconditioning strategy is presented to improve the parallel scalability.The results obtained for a radiation hydrodynamics computation within an ICF simulation demonstrate the efficiency and improvement of the adaptive strategy.For a typical model,the new strategy improves the parallel efficiency from 47% to 61%,and reduces the CPU time from 19.7 h to 14.5 h.
作者 徐小文 莫则尧 安恒斌 Xiaowen XU1,2, Zeyao MO1,2 , Hengbin AN1,2 (1 Laboratory of Computational Physics, Institute of Applied Physics and Computational Mathematics, Beijing 100094, China; 2 CAEP Software Center for High Performance Numerical Simulation, Beijing 100086, China)
出处 《中国科学:信息科学》 CSCD 北大核心 2016年第10期1411-1420,共10页 Scientia Sinica Informationis
基金 国家重点基础研究发展计划(973)(批准号:2011CB309702) 国家自然科学基金(批准号:61370067,91430218,91530324)资助项目 致谢 本文数值模拟应用测试是在北京应用物理与计算数学研究所激光聚变研究室开发的ICF内爆不稳定性模拟程序LARED-S上完成的,在测试过程中,程序研发团队成员范征锋研究员等为本文提供了测试模型,在此向他们表示感谢.
关键词 稀疏线性解法器 迭代方法 预条件子 代数多重网格算法(AMG) 并行计算 sparse linear solver iterative methods preconditioner algebraic multigrid(AMG) parallel computing
作者简介 通信作者.E-mail:xwxu@iapcm.ac.cnXiaowen XU was born in 1978. He received his B.S degree in computational mathematics from Xiangtan University (2002), and tlis Ph.D. degree in computational mathematics from the Graduate School of CAEP (2007), both in China. Currently, he is a professor of IAPCM. His research interests include parallel algorithms, scalable linear solvers, and large-scale simulations for real applications. Zeyao MO was born in 1971. He received his B.S in applied mathematics (1992) and Ph.D. in computer science (1997) both from National Defense University of Technology in China. Currently, he is a professor of IAPCM. His research interests include parallel computing, and parallel programming frameworks for large-scale numerical simulations.
  • 相关文献

参考文献30

  • 1Brandt A, McCormick S, Ruge J. Algebraic multigrid for sparse matrix equations. In: Sparsity and its Application. Cambridge: Cambridge University Press, 1984. 257-284. 被引量:1
  • 2Ruge J W, Stueben K. Algebraic multigrid. In: Multigrid Methods. Philadelphia: SIAM, 1987. 73-130. 被引量:1
  • 3Stueben K. Algebraic Multigrid (AMG): An Introduction With Applications. Sankt Augustin: GMD- Forschungszentrum Informationstechnik, 1999. 被引量:1
  • 4Yang U M. Parallel algebraic multigrid methods-high performance preconditioners. In: Numerical Solution of Partial Differential Equations on Parallel Computers. Berlin: Springer-Verlag, 2006. 209-236. 被引量:1
  • 5徐小文,莫则尧.并行代数多重网格算法可扩展性能分析[J].计算物理,2007,24(4):387-394. 被引量:6
  • 6Mo Z Y, Xu X W. Relaxed RSO or CLJP coarsening strategy for parallel AMG methods. Parall Comput, 2007, 33: 174-185. 被引量:1
  • 7Baker A H, Gamblin T, Schulz M, et al. Challenges of scaling AMG across modern multicore architectures. In: Proceedings of the 25th IEEE International Symposium on Parallel and Distributed Processing (IPDPS), Anchorage, 2011. 275-286. 被引量:1
  • 8Baker A H, Falgout R D, Gamblin T, et al. Scaling AMG solvers: on the road to Exascale. In: Proceedings of International Conference on Competence in High Performance Computing, Schwetzingen, 2010. 215-226. 被引量:1
  • 9Baker A H, Falgout R D, Gahvari H, et al. Preparing Algebraic Multigrid for Exascale. LLNL Technical Report LLNL-TR-533076. 2012. 被引量:1
  • 10Falgout R, Schroder J. Non-Galerkin coarse grids for algebraic multigrid. SIAM J Sci Comput, 2014, 36: 309-334. 被引量:1

二级参考文献158

  • 1徐小文 ,莫则尧 .一种新的并行代数多重网格粗化算法[J].计算数学,2005,27(3):325-336. 被引量:5
  • 2莫则尧,张爱清,曹小林,左风丽.多介质辐射流体力学数值模拟中的并行计算研究[J].自然科学进展,2006,16(3):287-292. 被引量:7
  • 3徐小文,莫则尧.并行代数多重网格算法可扩展性能分析[J].计算物理,2007,24(4):387-394. 被引量:6
  • 4Fu S W, Fu H Q, Shen L J, Huang S K, Chen G N. A nine point difference scheme and iteration solving method for two-dimensional energy equations with three temperatures[J]. Chinese J Comput Phys, 1998, 15(4): 489- 497. 被引量:1
  • 5Mo Z Y, Fu S W, Shen L J. Parallel numerical simulation for 2-dimension three temperature hydrodynamics[ J ]. Chinese J Comput Phys, 2000, 17(6): 625- 632. 被引量:1
  • 6Mo Z Y, Shen L J, Wittum G. Parallel adaptive muhigrid algorithm for 2-D 3-T diffusion equations[J]. Inter J Comput Math, 2004, 81(3) : 361 - 374. 被引量:1
  • 7Gu T X, Dai Z H, Hang X D, Fu S W, Liu X P. High efficient algebraic solution methods for 2D energy system of equations with three temperature[J]. Chinese J Comput Phys, 2005, 22(6) : 471 - 478. 被引量:1
  • 8Wu J P, Liu X P, Wang Z H, Dai Z H, Li X M. Two preconditioning techniques for two-dimension three-temperature energy equations[J]. Chinese J Comput Phys, 2005, 22(4) : 283 - 291. 被引量:1
  • 9Brandt A. Algebraic multigrid theory:the symmetric case[J]. App Math Comp, 1986,19:23 - 56. 被引量:1
  • 10Chang Q S, Wong Y S, Fu H. On the algebraic multigrid method[J]. JCP, 1996, 125: 279- 292. 被引量:1

共引文献35

投稿分析

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部 意见反馈