自由表面水流数值模拟程序并行化的预处理  被引量:2

Parallel Pretreatment of Free Surface Flow Numerical Simulation Program

在线阅读下载全文

作  者:李兴 魏清福 张景新[1,2] Li Xing;Wei Qingfu;Zhang Jingxin(School of Naval Architecture,Ocean&Civil Engineering,Shanghai Jiao Tong University,Shanghai200240,China;MOE Key Laboratory of Hydrodynamics,Shanghai Jiao Tong University,Shanghai200240,China;Jiangsu Taihu Planning and Design Institute of Water Resources Co.,Ltd.Shanghai Department,Shanghai200434,China)

机构地区:[1]上海交通大学船舶海洋与建筑工程学院,上海200240 [2]上海交通大学水动力学教育部重点实验室,上海200240 [3]江苏省太湖水利规划设计研究院有限公司上海分公司,上海200434

出  处:《水动力学研究与进展(A辑)》2023年第2期222-231,共10页Chinese Journal of Hydrodynamics

基  金:国家重点研发计划(2019YFB1503700)。

摘  要:并行化计算作为一种提升计算速度的手段被广泛应用于计算流体力学领域。为提升模拟速度,以及针对天然水域超大空间和时间尺度的CFD数值模拟需求,实现多节点分布式并行化计算很有必要。MPI作为目前较为成熟的并行接口是软件并行化的优先选择,其作为一种消息传递接口,通过通信方式实现程序的分布式并行。并行化过程中的并行分析、程序预处理以及通信模式的设计尤为必要。该文通过对自主开发的模拟软件的并行化分析,对预处理程序进行了开发,并设计了可扩展的通信模式,方便程序后续开发。通过分别执行串行程序与多线程并行程序模拟孤立波运动过程,分析并行程序计算的准确性,结果表明:并行计算程序与串行程序计算结果无显著差异,计算精度在10-6量级,且线程数不影响结果。Parallelization is widely used in the field of computational fluid dynamics to improve computational speed.In order to achieve more efficient simulation,especially for CFD applications to natural river flows with a huge spatial and temporal scale,it is necessary to realize multi-node distributed parallelization.MPI(message passing interface),as a relatively mature parallel interface,is the preferred choice for the codes paralleling.MPI achieves distributed parallelization of programs by means of the information communication.For the implementation of the program parallelization,the process of parallel analysis,pretreatment,and the design of the communication mode should be analysed in details.In this paper,the pretreatment program is developed based on the paralleling analysis of the codes.An extensible communication mode is designed to facilitate the subsequent development of the program.By simulating the solitary wave propagation in the serial program and the parallel program with multi-threads respectively,the calculation accuracy of parallel program is analyzed.The results show that there is no significant difference between the calculation results of parallel program and serial program,the calculation accuracy is in the order of 10^(-6),and the number of threads does not affect the results.

关 键 词:MPI 并行计算预处理 数据通讯 自由表面水动力学模拟 

分 类 号:TV139.2[水利工程—水力学及河流动力学]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象