面向结构动力学计算的撕裂有限元方法异构并行优化  

HETEROGENEOUS PARALLEL OPTIMIZATION OF TEARING FINITE ELEMENT METHODS FOR STRUCTURAL DYNAMICS CALCULATIONS

作　　者：聂宁明[1] 姚柯寒 曾艳 冯仰德[1,2] 王珏 李顺德[1] 张纪林 万健 林克豪[2] 高岳[3] 王彦棡 王宗国 Nie Ningming;Yao Kehan;Zeng Yan;Feng Yangde;Wang Jue;Li Shunde;Zhang Jilin;Wan Jian;Lin Kehao;Gao Yue;Wang Yangang;Wang Zongguo(Computer Network Information Center,Beijing 100190,China;College of Computer Science,Hangzhou Dianzi University,Hangzhou 310018,China;China Institute of Atomic Energy,Beijing 102413,China)

机构地区：[1]中国科学院计算机网络信息中心,北京100190 [2]杭州电子科技大学计算机学院,杭州310018 [3]中国原子能科学研究院,北京102413

出　　处：《数值计算与计算机应用》2024年第2期115-135,共21页Journal on Numerical Methods and Computer Applications

基　　金：国家重点研发计划(2020YFB0204802);中国科学院战略性先导科技专项(XDB0500103)资助.

摘　　要：本文结合大规模撕裂有限元方法和Newmark积分法,对结构动力学问题进行高精细的大规模并行求解.面向异构平台,设计了结点间和结点内的多级动静结合的负载均衡策略.在结点间,根据撕裂有限元方法划分子域边界特点,采用域边界平衡的图二分算法,均衡各个子域的计算量;在结点内,根据异构平台计算单元的性能差异,进行了计算负载的动态优化.针对核心计算模块批量矩阵向量乘进行多流并行优化,提升面向异构计算平台的利用率.本文优化已经集成到结构力学高性能数值模拟软件HARSA-feti中,实验采用真实反应堆核燃料组件的流致振动仿真作为算例,结果表明模拟性能提高了71.3%以上,首次实现了百亿网格规模的全堆芯燃料棒组件的高精细模拟,相较于1000块GPU,16000块GPU的强、弱可扩展并行效率分别达到74.1%和81.1%.This paper adopts the Newmark integration method based on the large-scale tearing finite element method to perform high-precision large-scale parallel solving of structural dynamic calculations.A multi-level load balancing strategy combining static and dynamic methods is designed for heterogeneous platforms.For inter-node computing,subdomain boundaries are partitioned based on the characteristics of the tearing finite element method,and a domain boundary balanced graph bipartition algorithm is used to balance the computation load of each subdomain.For intra-node computing,dynamic optimization of computation load is performed based on the performance differences of computing units on heterogeneous platforms.To improve the utilization rate of heterogeneous computing platforms,multi-stream parallel optimization is carried out for the core computing module's batch matrix-vector multiplication.The optimization in this paper has been integrated into the high-performance numerical simulation software for structural mechanics,HARSA-feti.The simulation performance is demonstrated using the flow-induced vibration simulation of a real reactor fuel component as an example.The results show that the simulation performance has increased by more than 71.3\%,and the high-precision simulation of a billion-grid-scale full-core fuel rod component has been achieved for the first time.Compared with 1,000 GPUs,the strong and weak scalable parallel efficiency of 16,000 GPUs reached 74.1\%and 81.1\%,respectively.

关 键 词：结构动力学 大规模并行计算 负载均衡 异构计算 矩阵向量乘 

分 类 号：O241.82[理学—计算数学] O302[理学—数学]