第一性原理软件北京原子技术模拟工具包的重构进展

马英晋; 张坦; 何连花; 金钟

doi:10.7498/aps.69.20191658

摘要

基于第一性原理的理论方法的研究, 代表了材料计算、分子模拟等领域的科学高地, 相应的第一性原理计算软件直接关系到该领域相关理论、算法的积累. 本文汇报了我们在重构第一性原理计算模拟软件—北京原子技术模拟工具包(BSTATE)的一些最新进展. 重构的核心思想是降低用户使用门槛、扩展软件适用范围、增加软件对于流行计算框架的支持. 基于此思路, 在BSTATE原有Makefile编译系统的基础上添加了CMake编译环境, 并支持各种数学函数库的自动和交互式配置; 通过在原有内置泛函基础上增添Libxc泛函库的支持, 使BSTATE支持的泛函数量有了数量级上的增长; 分析测试BSTATE在集群的并行特点, 并以更新数学库接口(FFTW3、Cufftw)的形式提供对于流行异构框架的初步支持.

关键词:

Abstract

The development of first principle methods can represent the summit of the sciences in the material computing and molecular modeling, and the corresponding first principle software packages are closely related with the accumulation of theories and algorithms in this field. In this paper, we reported our recent progress in refactoring the first principle package BSTATE. The key points in the reconstruction are lowering the doorsill, extending the scope of application, as well as adjusting package to the popular computer hardware. And as such, we updated the Makefile system to the new CMake system, in which the GUI can be used and many math libraries can be configured automatically; we added the support for the Libxc library, in which a large quantity of density functionals are included; we updated the interface for supporting GPU, in order to support the heterogeneous computing system. After refactoring, the Makefile system of BSTATE can supply both the Makefile and CMake system, and the Fourier transform libraries such as FFTW2, FFTW3, and Cufftw, the math libraries such as Intel MKL library, Openblas, and the density functional library such as Libxc, can be automatically or manually assigned. The integration of FFTW3 can slightly prompt the calculating efficiency in Intel’s many integrated core (MIC) architecture, and the integration of Cufftw can supply the initial support for the graphics processing unit (GPU) architecture, respectively. The usage of Libxc library makes the BSTATE package has the capacity to use hundreds density functionals, and the usages of various functionals were demonstrated by calculating the density of states of GaAs compound. Beyond the integration of various libraries, the parallel performance of BSTATE was also investigated. It can be found that the Fourier transformation and the solving for the eigenvalue equations are the major contributions. Using the tuning and analysis utilities (TAU) tool, we found that the tasks can be well distributed in modern HPC clusters. It implied that the refactoring didn’t affect the parallel efficiency of original BSTATE package. In a following benchmark test of graphene fragments, one can found that the refactored BSTATE package showed the best performance, its FFTW3 & Libxc version owns about 0–17% acceleration comparing to that of FFTW2 version.

Keywords:

first principle computational software /
density-functional theory /
Beijing Simulation Tool for Atom TEchnology /
Software refactoring

作者及机构信息

1.
中国科学院计算机网络信息中心, 北京　100190

2.
中国科学院计算科学应用研究中心, 北京　100190

3.
北京凝聚态物理国家研究中心, 中国科学院物理研究所, 北京　100190

4.
中国科学院大学物理科学学院, 北京　100049

通信作者: 何连花, helh@sccas.cn

基金项目: 中科院-中国科学院信息化专项-第一性原理计算模拟软件(XXH13506-403)

Authors and contacts

1.
Computer Network Information Center, Chinese Academy of Science, Beijing 100190, China

2.
Center of Scientific Computing Applications & Research, Chinese Academy of Science, Beijing 100190, China

3.
Beijing National Research Center for Condensed Matter Physics, Institute of Physics, Chinese Academy of Science, Beijing 100190, China

4.
School of Physical Sciences, University of Chinese Academy of Sciences, Beijing 100049, China

Corresponding author: He Lian-Hua, helh@sccas.cn

文章全文

参考文献

[1]	Frisch M J, Trucks G W, Schlegel H B, et al. 2016 Gaussian Inc. Wallingford CT
[2]	Kresse G, Furthmüller J 1996 Comp. Mater. Sci. 6 15 Google Scholar
[3]	Liu W, Wang F, Li L 2003 J. Theor. Comput. Chem. 2 257 Google Scholar
[4]	Li P, Liu X, Chen M, Lin P, Ren X, Lin L, Yang C, He L 2016 Comput. Mater. Sci. 112 503 Google Scholar
[5]	Fang Z, Terakura K J 2002 Phys. Condens. Mat. 14 3001 Google Scholar
[6]	Troullier N, Martins J L 1991 Phys. Rev. B 43 1993 Google Scholar
[7]	Vanderbilt D 1990 Phys. Rev. B 41 7892 Google Scholar
[8]	Kresse G, Hafner J J 1994 Phys.: Condens. Matter 6 8245 Google Scholar
[9]	Deng X Y, Wang L, Dai X, Fang Z 2009 Phys. Rev. B 79 075114 Google Scholar
[10]	Rullier-Albenque F, Alloul H, Balakirev F, Proust C 2008 Europhys. Lett. 81 37008 Google Scholar
[11]	Rice T M, Ueda K 1985 Phys. Rev. Lett. 55 995 Google Scholar
[12]	Yu R, Zhang W, Zhang H J, Zhang S C, Dai X, Fang Z 2010 Science 329 61 Google Scholar
[13]	余睿, 张薇, 翁红明, 戴希, 方忠 2010 物理 39 618 Yu R, Zhang W, Weng H M, Dai X, Fang Z 2010 Physics 39 618
[14]	梁拥成, 张英, 郭万林, 姚裕贵, 方忠 2007 物理 36 385 Google Scholar Liang Y C, Zhang Y, Guo W L, Yao Y G, Fang Z 2007 Physics 36 385 Google Scholar
[15]	徐刚, 戴希, 方忠 2009 物理 38 651 Google Scholar Xu G, Dai X, Fang Z 2009 Physics 38 651 Google Scholar
[16]	Dirac P A M 1930 Proc. Camb. Phil. Soc. 26 376 Google Scholar
[17]	Perdew J P, Chevary J A, Vosko S H, Jackson K A, Pederson M R, Singh D J, Fiolhais C 1992 Phys. Rev. B 46 6671 Google Scholar
[18]	Perdew J P, Burke K, Ernzerhof M 1997 Phys. Rev. Lett. 77 3865
[19]	Kohn W, Sham L J 1965 Phys. Rev. 140 A1133 Google Scholar
[20]	Hohenberg P, Kohn W 1964 Phys. Rev. 136 B864 Google Scholar
[21]	黄美纯 2000 物理学进展 20 199 Google Scholar Huang M C 2000 Progress in Physics 20 199 Google Scholar
[22]	李震宇, 贺伟, 杨金龙 2005 化学进展 17 192 Google Scholar Li Z Y, H W, Yang J L 2005 Progress In Chemistry 17 192 Google Scholar
[23]	Heyd J, Scuseria G E 2003 J. Chem. Phys 118 8207 Google Scholar
[24]	Heyd J, Scuseria G E 2006 J. Chem. Phys 124 219906 Google Scholar
[25]	Clemencic M, Mato P 2012 J. Phys.: Conf. Ser. 396 052021 Google Scholar
[26]	Frigo M, Johnson S G 2005 Proceedings of the IEEE 93 216 Google Scholar
[27]	Lehtola S, Steigemann C, Oliveira M J T, Marques M A L 2018 Software X 7 1
[28]	Engel E, Keller S, Bonetti A F, Müller H, Dreizler R M 1995 Phys. Rev. A. 52 2750 Google Scholar
[29]	Rae A 1973 Chem. Phys. Lett. 18 574 Google Scholar

施引文献

图 1 BSTATE计算流程示意图

Fig. 1. Flowchart of BSTATE package.

下载: 全尺寸图片幻灯片

图 2 闪锌矿结构的GaAs双原子化合物

Fig. 2. GaAs compound with zinc blende structure.

下载: 全尺寸图片幻灯片

图 3 多种泛函计算的GaAs的态密度 (a) (b) (c) (d)

Fig. 3. Calculated density of states (DOS): (a); (b); (c); (d)

下载: 全尺寸图片幻灯片

图 4 个人PC (左图)及集群MPI (右图)测试中BSTATE各模块的时间占比统计

Fig. 4. Time percentages of subroutines in BSTATE using PC (left) and Cluster (right).

下载: 全尺寸图片幻灯片

图 5 BSTATE的MPI多线程并行分析; 其中蓝色、红色、黄色、黑色分别代表application(BSTATE), MPI_barrier, MPI_reduce以及MPI_allreduce.

Fig. 5. Analysis via TAU performance system framework for BSTATE

下载: 全尺寸图片幻灯片

图 6 重构前后的BSTATE跨节点并行测试

Fig. 6. Benchmark of MPI parallel for BSTATE with and without refactoring.

下载: 全尺寸图片幻灯片

表 1 重构前后BSTATE编译系统的对比

Table 1. Comparison of BSTATE compilation system.

项目	重构前	重构后
编译系统	GNUMake	CMake
图形GUI	不支持	支持
跨平台	手工修改Makefile文件提供支持	原生支持
数学库	手工配置	自动配置
外置函数库	手工配置	支持自动配置
异构支持	否	是
高级编译选项	手工配置	支持GUI配置
多线程编译	不支持	支持
用户门槛	高	低

下载: 导出CSV

表 2 Libxc版本与原始版本性能比较

Table 2. Benchmarks between BSTATEs with/without Libxc.

项目	单核/s	多核/s
BSTATE	42.4	22.6
BSTATE+Libxc	43.6	23.4
性能比	0.97	0.97
^* 测试机器为AMD A10 PRO-7800 B R7 (4核); GaAs体系

下载: 导出CSV

表 3 FFTW3+Libxc版本与原始FFTW2版本性能比较

Table 3. Benchmarks between V.FFTW2 and V.FFTW3+ Libxc.

项目	CPU平台/s	MIC平台/s
FFTW2	1181	1717
FFTW3+Libxc	1179	1593
性能比	1.00	1.08
^* CPU平台为Intel至强E7-4830v3 (56核); 石墨烯体系 ^* MIC平台为Intel Phi-7210 (64核); 石墨烯体系

下载: 导出CSV

map

[1]	Frisch M J, Trucks G W, Schlegel H B, et al. 2016 Gaussian Inc. Wallingford CT
[2]	Kresse G, Furthmüller J 1996 Comp. Mater. Sci. 6 15 Google Scholar
[3]	Liu W, Wang F, Li L 2003 J. Theor. Comput. Chem. 2 257 Google Scholar
[4]	Li P, Liu X, Chen M, Lin P, Ren X, Lin L, Yang C, He L 2016 Comput. Mater. Sci. 112 503 Google Scholar
[5]	Fang Z, Terakura K J 2002 Phys. Condens. Mat. 14 3001 Google Scholar
[6]	Troullier N, Martins J L 1991 Phys. Rev. B 43 1993 Google Scholar
[7]	Vanderbilt D 1990 Phys. Rev. B 41 7892 Google Scholar
[8]	Kresse G, Hafner J J 1994 Phys.: Condens. Matter 6 8245 Google Scholar
[9]	Deng X Y, Wang L, Dai X, Fang Z 2009 Phys. Rev. B 79 075114 Google Scholar
[10]	Rullier-Albenque F, Alloul H, Balakirev F, Proust C 2008 Europhys. Lett. 81 37008 Google Scholar
[11]	Rice T M, Ueda K 1985 Phys. Rev. Lett. 55 995 Google Scholar
[12]	Yu R, Zhang W, Zhang H J, Zhang S C, Dai X, Fang Z 2010 Science 329 61 Google Scholar
[13]	余睿, 张薇, 翁红明, 戴希, 方忠 2010 物理 39 618 Yu R, Zhang W, Weng H M, Dai X, Fang Z 2010 Physics 39 618
[14]	梁拥成, 张英, 郭万林, 姚裕贵, 方忠 2007 物理 36 385 Google Scholar Liang Y C, Zhang Y, Guo W L, Yao Y G, Fang Z 2007 Physics 36 385 Google Scholar
[15]	徐刚, 戴希, 方忠 2009 物理 38 651 Google Scholar Xu G, Dai X, Fang Z 2009 Physics 38 651 Google Scholar
[16]	Dirac P A M 1930 Proc. Camb. Phil. Soc. 26 376 Google Scholar
[17]	Perdew J P, Chevary J A, Vosko S H, Jackson K A, Pederson M R, Singh D J, Fiolhais C 1992 Phys. Rev. B 46 6671 Google Scholar
[18]	Perdew J P, Burke K, Ernzerhof M 1997 Phys. Rev. Lett. 77 3865
[19]	Kohn W, Sham L J 1965 Phys. Rev. 140 A1133 Google Scholar
[20]	Hohenberg P, Kohn W 1964 Phys. Rev. 136 B864 Google Scholar
[21]	黄美纯 2000 物理学进展 20 199 Google Scholar Huang M C 2000 Progress in Physics 20 199 Google Scholar
[22]	李震宇, 贺伟, 杨金龙 2005 化学进展 17 192 Google Scholar Li Z Y, H W, Yang J L 2005 Progress In Chemistry 17 192 Google Scholar
[23]	Heyd J, Scuseria G E 2003 J. Chem. Phys 118 8207 Google Scholar
[24]	Heyd J, Scuseria G E 2006 J. Chem. Phys 124 219906 Google Scholar
[25]	Clemencic M, Mato P 2012 J. Phys.: Conf. Ser. 396 052021 Google Scholar
[26]	Frigo M, Johnson S G 2005 Proceedings of the IEEE 93 216 Google Scholar
[27]	Lehtola S, Steigemann C, Oliveira M J T, Marques M A L 2018 Software X 7 1
[28]	Engel E, Keller S, Bonetti A F, Müller H, Dreizler R M 1995 Phys. Rev. A. 52 2750 Google Scholar
[29]	Rae A 1973 Chem. Phys. Lett. 18 574 Google Scholar

计量

文章访问数: 12475
PDF下载量: 182
被引次数: 0

姓名
邮箱
手机号码
标题
留言内容
验证码

搜索

留言板