GPU集群下第一原理非局部映射势能计算

doi:10.3724/SP.J.1087.2013.01540

计算机应用 ›› 2013, Vol. 33 ›› Issue (06): 1540-1552.DOI: 10.3724/SP.J.1087.2013.01540

GPU集群下第一原理非局部映射势能计算

付继芸¹,²,贾伟乐¹,²,曹宗雁¹,王龙¹,叶煌¹,迟学斌¹

1. 中国科学院计算机网络信息中心,北京 100190
2. 中国科学院研究生院,北京 100190

收稿日期:2012-12-27 修回日期:2013-02-22 发布日期:2013-06-05 出版日期:2013-06-01
通讯作者: 付继芸
作者简介:付继芸(1985-)，女，河南安阳人，硕士研究生，主要研究方向：计算机软件与理论、并行计算、高性能计算；贾伟乐(1985-)，男，河北衡水人，硕士研究生，主要研究方向：计算机软件与理论、并行计算、高性能计算；曹宗雁(1981-)，男，湖南湘潭人，博士研究生，主要研究方向：高性能计算机系统。
基金资助:
国家自然科学基金资助项目（61202054）;国家863计划项目（2010AA012301）;中国科学院知识创新工程项目（CNIC_ZR_201202）;中国科学院“十二五”信息化专项

First-principle nonlocal projector potential calculation on GPU cluster

FU Jiyun¹,²,JIA Weile¹,²,CAO Zongyan¹,WANG Long¹,YE Huang¹,CHI Xuebin¹

1. Computer Network Information Center, Chinese Academy of Sciences, Beijing 100190, China
2. Graduate School of Chinese Academy of Sciences, Beijing 100190, China

Received:2012-12-27 Revised:2013-02-22 Online:2013-06-05 Published:2013-06-01
Contact: FU Jiyun

摘要/Abstract

摘要： 平面波赝势密度泛函（PWP-DFT）计算是材料计算中应用最广泛的方法,其中映射计算是PWP-DFT方法求解自洽迭代中重要的一部分。针对映射势能计算成为软件加速的瓶颈，提出了针对该部分的图形处理器(GPU)加速算法，其中考虑GPU的特点:1）使用了新的并行机制求解非局部映射势能；2）重新设计了数据分布结构；3）减少内存的使用；4）提出了一种解决算法中数据相关问题的方法。最终获得了18~57倍加速，使每步分子动力学模拟最终降为12s。详细分析了该模块在GPU平台上的测试时间，同时对该算法在GPU集群上的计算瓶颈进行了讨论。

关键词: 第一性原理, 密度泛函理论, 赝势平面波, 非局部映射势能, GPU加速

Abstract: Plane Wave Pseudopotential (PWP) Density Functional Theory (DFT) calculation is the most widely used method for material calculation. The projector calculation plays an important part in PWP-DFT calculation for the self-consistent iteration solution, while it often becomes a hinder to the speed-up of software. Therefore, according to the features of Graphic Processing Unit (GPU), a speed-up algorithm was proposed: 1) using a new parallel mechanism to solve the potential energy of nonlocal projector, 2) redesigning the distribution structure of data, 3) reducing the use of computer memory, 4) Proposing a solution to the related data problems of the algorithm. Eventually got 18-57 times acceleration, and reached the 12 seconds per step of the molecular dynamics simulation. In this paper, the testing time of running this model on GPU platform was analysed in detail, meanwhile the calculation bottleneck of the implementation of this method into GPU clusters was discussed

Key words: first-principle, Density Functional Theory (DFT), Plane Wave Pseudopotential (PWP), nonlocal projector potential, GPU speedup

中图分类号:

付继芸贾伟乐曹宗雁王龙叶煌迟学斌. GPU集群下第一原理非局部映射势能计算[J]. 计算机应用, 2013, 33(06): 1540-1552.

FU Jiyun JIA Weile CAO Zongyan WANG Long YE Huang CHI Xuebin. First-principle nonlocal projector potential calculation on GPU cluster[J]. Journal of Computer Applications, 2013, 33(06): 1540-1552.

参考文献

［1］薛定谔方程［EB/OL］.［2012-11-30］.http://zh.wikipedia.org/wiki/%E8%96%9B%E5%AE%9A%E8%B0%94%E6%96%B9%E7%A8%8B

［2］PAYNE M C, TETER M P, ALLAN D C, et al. Iterative minimization techniques for ab initio total-energy calculations: molecular dynamics and conjugate gradients［J］.Reviews of Modern Physics,1992, 64(4):1045-1097.

［3］CARTER E A. Challenges in modeling materials properties without experimental Input［J］. Science, 2008, 321(5890): 800-803.

［4］KENT P R C. Computational challenges in nanoscience: An ab initio perspective［EB/OL］.［2012-11-30］. http://computing.ornl.gov/workshops/peta08/presentations/p_kent.pdf.

［5］WANG L W, LEE B, SHAN Z H, et al. Linear scaling 3D fragment method for large scale electronic structure calculations ［C］// Proceedings of the 2008 ACM/IEEE Conference on Supercomputing. Piscataway:IEEE, 2008:65.

［6］WANG L W. PEtot code ［EB/OL］.［2012-11-30］.https://hpcrd.lbl.gov/~linwang/PEtot/PEtot.html.

［7］TOMONO H, AOKI M, IITAKA T, et al. GPU based acceleration of first principles calculation［J］.Joural of Physics,2010,215(1):012121.

［8］GENOVESE L, OSPICI M, DEUTSCH T, et al. Density functional theory calculation on many-cores hybrid CPU-GPU architectures［J］.Journal Chemistry Physics,2009(131):034103.

［9］WANG L, JIA W L, CHI X B, et al. Large scale plane wave pseudopotential density function theory calculations on GPU clusters ［C］// Proceedings of the 2011 International Conference for High Performance Computing, Networking, Storage and Analysis. New York: ACM, 2011:71.

［10］JIA W L, CAO Z Y, WANG L, et al. The analysis of a plane wave pseudopotential density functional theory code on a GPU machine［J］. Computer Physics Communications, 2013,184(1):9-18.

［11］WANG L W. Large scale LDA-band-gap-corrected GaAsN calculations［J］. Applied Physics Letters,2001，78（11): 1565.

［12］TOP500 Supercomputer Sites［EB/OL］.［2012-11-30］. http://www.top500.org/

［13］巴塞罗那超级计算中心将部署全球首个基于ARM的CPU/GPU混合型超级计算机［EB/OL］.［2012-11-30］.http://tech.163.com/digi/11/1116/11/7IVS4OR300162OUT.html.

[1]	姜涛, 梁振宇, 程然, 金耀初. GPU加速的演化算法求解多目标流水车间调度问题[J]. 《计算机应用》唯一官方网站, 2024, 44(5): 1364-1371.
[2]	覃方涛房斌. GPU加速的二值图连通域标记并行算法[J]. 计算机应用, 2010, 30(10): 2774-2776.

GPU集群下第一原理非局部映射势能计算

First-principle nonlocal projector potential calculation on GPU cluster

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 2

编辑推荐

Metrics