Application and Development of MPI in Monte Carlo Code GMT
-
摘要: 针对ADS颗粒靶概念的研究和设计,中国科学院近代物理研究所自主研发了蒙特卡罗模拟软件GMT。为了提高GMT程序的计算效率,研究了MPI在GMT中的应用和发展,实现了大规模随机数在进程中的随机分配,并采用快速读写文件的方式替代了MPI相关数据通信函数,极大地提高了计算效率。并研究了不同规模计算实例进程数、加速比、效率之间的关系,确定了最大加速进程数及并行效率最高时的进程数,为科研工作者在计算资源和计算效率之间选择最优计算方案提供了科学依据。MPI在GMT中的成功应用使计算资源得到了充分、高效的利用,极大地提高了计算效率,解决了蒙特卡罗方法中大规模事件模拟计算时间长、计算不稳定等问题,在散裂靶大规模扫描计算中发挥了重要的作用。
For the research and design of the ADS granular-flow target concept, the Institute of Modern Physics, CAS has developed a Monte Carlo simulation software (GPU-accelerated Monte Carlo Transport program, GMT). In order to improve the computational efficiency of the GMT program, development and application of MPI in GMT were studied, to realize random distribution of the large-scale random number in the sub processes. Rapid reading and writing files were employed instead of the MPI data communication function, which greatly improves the computational efficiency. Different scale calculations were performed to study the relationship of process instance number, speedup to find the maximum acceleration process number and the number of processes when parallel efficiency is highest, which provides a scientific basis for researchers to optimize the computational program between computational resources and computation efficiency. The successful application of MPI in GMT, utilizes the computing resources fully and efficiently, improves the computational efficiency, solve the long time cost and unstable problem of Monte Carlo method in large-scale event simulations, plays an important role in the large-scale scanning calculation of the spallation target.Abstract: For the research and design of the ADS granular-flow target concept, the Institute of Modern Physics, CAS has developed a Monte Carlo simulation software (GPU-accelerated Monte Carlo Transport program, GMT). In order to improve the computational efficiency of the GMT program, development and application of MPI in GMT were studied, to realize random distribution of the large-scale random number in the sub processes. Rapid reading and writing files were employed instead of the MPI data communication function, which greatly improves the computational efficiency. Different scale calculations were performed to study the relationship of process instance number, speedup to find the maximum acceleration process number and the number of processes when parallel efficiency is highest, which provides a scientific basis for researchers to optimize the computational program between computational resources and computation efficiency. The successful application of MPI in GMT, utilizes the computing resources fully and efficiently, improves the computational efficiency, solve the long time cost and unstable problem of Monte Carlo method in large-scale event simulations, plays an important role in the large-scale scanning calculation of the spallation target.-
Key words:
- ADS granular-flow target /
- MPI /
- GMT /
- random number /
- data transmission /
- speedup
-
[1] ZHAN Wenlong, XU Hushan. Bulletin of National Academy of Sciences, 2012, 27(3): 375. (in Chinese)(詹文龙, 徐瑚珊. 中国科学院院刊, 2012, 27(3): 375) [2] YANG Lei, ZHAN Wenlong. Science China Technological Sciences, 2015, 58: 1. [3] DENISE B, PELOWIT Z. MCNPXTM Users's manual. Ver-sion 2.6.0, LA-CP-07-1473. US, Los Alamos National Labo-ratory, 2008: 1. [4] AGOSTINELLI S, ALLISON J, AMAKO K, et al. Nucl Instr and Meth A, 2003, 506(3): 250. [5] ALFREDO F, PAOLA R S, ALBERTO F, et al. FLUKA:a Multi-particle Transport Code (Program version2005), in:CERN 2005-10 (2005), INFN/TC 05/11, SLAC-R-773. [6] DENG Li, LI Gang. Chinese Journal of computational Physics, 2010, 27: 791. (in Chinese)(邓力, 李刚. 计算物理, 2010, 27: 791.) [7] WANG Yi, YANG Pingli, ZHU Weijie, et al. Nuclear Elec-tronics& Detection Technology, 2001, 21(1): 31. (in Chinese)(王义, 杨平利, 朱伟杰, 等.核电子学与探测技术, 2001, 21(1):31.) [8] WANG Lei, WANG Kan, YU Ganglin.Nuclear Electron-ics& Detection Technology, 2008, 28: 163.(in Chinese)(王磊, 王侃, 余纲林.核电子学与探测技术, 2008, 28: 163.) [9] LU Fengshun, SONG Junqiang, YING Fukang, et al. Com-puter Science, 2011, 38(3): 5. (in Chinese)(卢风顺, 宋君强, 银福康, 等. 计算机科学, 2011, 38(3): 5.) [10] YANG Bo. Research on CPU/GPU Synergetic Algorithm for Monte Carlo Deep Penetration Particle Transport[D]. Chang-shai: Graduate School of National University of Defense Technology, 2011. (in Chinese)(杨博.深穿透粒子输运蒙特卡罗模拟的CPU/GPU协同算法研究[D].长沙: 国防科学技术大学研究生院, 2011.) [11] DU Zhihui. High Performance Parallel programming-MPI Parallel Programming[M]. BeiJing: Tsinghua University Press, 2001. (in Chinese)(都志辉.高性能并行编程技术一MPI并行程序设计[M].北京:清华大学出版社, 2001.) [12] FORREST B. The MCNP5 Random Number Generator, LA-UR-07-7963. US, Los Alamos National Laboratory, 2002: 1.
计量
- 文章访问数: 1102
- HTML全文浏览量: 83
- PDF下载量: 119
- 被引次数: 0