计算机科学 ›› 2017, Vol. 44 ›› Issue (3): 32-35.doi: 10.11896/j.issn.1002-137X.2017.03.008

• 2015全国高性能计算学术年会 • 上一篇    下一篇

Calculix三级并行优化及其在天河二号超级计算机中的应用

姜文超,林穗,王多强,李东明,金海   

  1. 广东工业大学计算机学院 广州510006,广东工业大学计算机学院 广州510006,华中科技大学计算机学院 武汉430074,广州船舶及海洋工程设计研究院 广州510250,华中科技大学计算机学院 武汉430074
  • 出版日期:2018-11-13 发布日期:2018-11-13
  • 基金资助:
    本文受广东省科技计划项目(2015B010109001),广东省产学研合作项目(2015B090901051),教育部留学回国人员启动基金项目(14ZK0152),NSFC-广东联合基金(第二期)超级计算科学应用研究专项资助

Three-level Parallel Optimization and Application of Calculix in TH-2 Super-computing Environments

JIANG Wen-chao, LIN Sui, WANG Duo-qiang, LI Dong-ming and JIN Hai   

  • Online:2018-11-13 Published:2018-11-13

摘要: 针对开源有限元软件Calculix传统计算模式在大规模数值计算中的低效问题,提出了Calculix三级并行优化策略,即预处理并行优化、节点间并行调度以及节点内多核多线程并行改造。预处理并行优化在方程组分解过程中与分解过程后,分别对其参数矩阵进行有条件的动态舍弃,据此构造了部分列选主元多行双门槛不完全LU分解预处理算法,并对算法的可行性、有效性以及收敛性给出了证明。为充分发挥TH-2超级计算机强大的资源优势,相继给出了基于QoS的节点间任务动态调度算法,以及节点内多核多线程并行任务调度算法,进一步实现计算任务与资源之间的优化匹配和QoS需求。在实验环节中搭建了针对天河二号(TH-2)超级计算环境的有限元并行计算与分析平台,并完成了针对船舶疲劳强度分析问题的实际工程应用测试。理论分析与工程算例测试结果充分证明:Calculix三级并行优化方案能够有效提高Calculix求解线性方程组的速度,在可获取足够计算资源的前提下, 与传统计算模式相比 ,实际工程算例的计算速度平均提高了2~4倍。

关键词: 并行计算,超级计算,Calculix,并行优化,疲劳强度分析

Abstract: To increase the computing efficiency of Calculix in large scale numerical calculation,this paper researched Calculix and proposed a three-level parallel optimization scheme including preprocessing-level,task level and thread le-vel.A multiple row double threshold incomplete LU matrix decomposition strategy based on column pivoting and dyna-mic abandon was presented.The corresponding algorithm was given and proved to be efficient,available and convergent.Furthermore,to utilize the powerful computing resources of TH-2 supercomputer,a task level parallel scheduling algorithm using between computing nodes and a thread level parallel algorithm using between multiple computing cores were also developed and deployed in the experimental platform which focused on the ship fatigue analysis.Both the theory analysis and the actual engineering cases testing were provided to show that the three level parallel optimization scheme based on Calculix can increase the solving speed of liner equations and the analysis efficiency in engineering design areas,and the average speed-up ratio can reach about 2~4 when enough resources can be obtained.

Key words: Parallel computing,Super-computing,Calculix,Parallel optimization,Fatigue strength analysis

[1] GUO L L,CHEN Z F,LUO J R,et al.A Review of the Exten-ded Finite Element Method and its Applications[J].Chinese Quarterly of Mechanics,2011,2(4):612-625.(in Chinese) 郭历伦,陈忠富,罗景润,等.扩展有限元方法及应用综述[J].力学季刊,2011,32(4):612-625.
[2] GOURI D,EMMANUEL L.Finite Element Method[M].John Wiley & Sons,2012.
[3] YU Y T,DU P G,WANG Z W.Research on the current application status of finite element method[J].Journal of Machine Design,2005,2(3):6-9.(in Chinese) 于亚婷,杜平安,王振伟.有限元法的应用现状研究[J].机械设计,2005,2(3):6-9.
[4] CALCULIX.A Free Software Three-Dimensional Structural Finite Element Program .http://www.calculix.de.
[5] FUNG Y C.Foundations of Solid Mechanics [M].Englewood Cliffs,N J:Prentice Hall,1965.
[6] DHONDT G,WITTIG K.Calculix:A free software three-di-mensional structural finite element program.http:/www.calculix.de.
[7] WANG H,DING J H.Constructing ANSYS distributed highperformance platform using LINUX cluster technology[J].Development and Application of High Performance Computing,2011(1):34-39.(in Chinese) 王惠,丁峻宏.LINUX集群技术构建ANSYS分布式高性能计算平台[J].高性能计算发展与应用,2011(1):34-39.

No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!