基于多GPU的格子Boltzmann法对槽道湍流的直接数值模拟

许丁; 陈刚; 王娴; 李跃明

doi:10.3879/j.issn.1000-0887.2013.09.009

基于多GPU的格子Boltzmann法对槽道湍流的直接数值模拟

doi: 10.3879/j.issn.1000-0887.2013.09.009

西安交通大学航天航空学院, 机械结构强度与振动国家重点实验室,西安 710049

基金项目: 国家自然科学基金资助项目（11242010；11102150）；中央高校基本科研业务费专项资金资助项目

详细信息

作者简介:
许丁 (1980—)，男，博士，讲师(E-mail:dingxu@mail.xjtu.edu.cn);王娴 (1977—)，女，博士，副教授（通讯作者.E-mail: wangxian@mail.xjtu.edu.cn）.

中图分类号: TB126; O351
计量
- 文章访问数: 1796
- HTML全文浏览量: 180
- PDF下载量: 1392
- 被引次数: 0
出版历程
- 收稿日期: 2013-05-30
- 修回日期: 2013-06-05
- 刊出日期: 2013-09-15

Direct Numerical Simulation of the Wall-Bounded Turbulent Flow by Lattice Boltzmann Method Based on Multi-GPU

State Key Laboratory for Strength and Vibration of Mechanical Structures, School of Aerospace, Xi’an Jiaotong University, Xi’an 710049, P.R.China

Funds: The National Natural Science Foundation of China（11242010；11102150）

摘要

摘要: 采用多GPU并行的格子Boltzmann方法（lattice Boltzmann method, LBM）对充分发展的槽道湍流进行了直接数值模拟．GPU(graphic processing unit)的数据并行单指令多线程（single-instruction multiple-thread, SIMT)特征与LBM完美的并行性相匹配，使得LBM求解器在GPU上运行获得了极高的性能，亦使得大规模DNS(direct numerical simulation)在桌面级计算机上进行成为可能．采用8个GPU，网格数目达到6.7×10⁷，全场网格尺寸Δ⁺=1.41．模拟3×10⁶个时间步长，用时仅24 h．另外，直接模拟结果无论是在平均流速或湍流统计量上均与Moser等的结果吻合得很好，这也证实了二阶精度的格子Boltzmann法直接模拟湍流的能力与有效性
- 格子Boltzmann法 /
- 多GPU并行 /
- 槽道湍流 /
- 直接数值模拟(DNS)
Abstract: The wallbounded turbulent flow was simulated directly (DNS) by lattice Boltzmann method (LBM) through multi-GPU parallel computing. The Data-parallel SIMT (single-instruction multiple-thread) characteristic of GPU matched the parallelism of LBM well, which led to high efficiency of GPU on the LBM solver. At the same time, it bronght possibility for large-scale DNS on the desk-top supercomputer. In this DNS work, 8 GPUs were adopted. The number of meshes is 6.7×10⁷, which resulted in a non-dimensional mesh size of Δ⁺=1.41 for the whole solution domain. It took only 24 hours for the GPU-LBM solver to simulate 3×10⁶ LBM steps. As a result, both the mean velocity and turbulent variables, such as Reynolds stress and velocity fluctuations, agree well with the results of Moser, et al. The capacity and validity of LBM in simulating turbulent flow are verified.
- lattice Boltzmann method /
- multi-GPU parallel computing /
- wall-bounded turbulent flow /
- DNS

HTML全文

参考文献(19)

[1]	Chen S Y, Doolen G D. Lattice Boltzmann method for fluid flows[J]. Annual Review of Fluid Mechanics,1998,30: 329-364.
[2]	何雅玲, 王勇, 李庆. 格子Boltzmann方法的理论及应用[M]. 北京：科学出版社, 2009. (HE Ya-ling, WANG Yong, LI Qing.Lattice Boltzmann Method: Theory and Applications[M]. Beijing: Science Press, 2009. (in Chinese))
[3]	Yu H, Girimaji S S, Luo L S. DNS and LES of decaying isotropic turbulence with and without frame rotation using lattice Boltzmann method[J]. Journal of Computational Physics,2005,209(2): 599-616.
[4]	Yu H, Luo L S, Girimaji S S. LES of turbulent square jet flow using an MRT lattice Boltzmann model[J].Computers &Fluids,2006,35(8/9): 957-965.
[5]	Moser R D, Kim J, Mansour N N. Direct numerical simulation of turbulent channel flow up to Re_τ=590[J]. Phys Fluids,1999,11(4): 943-945.
[6]	Kim J, Moin P, Moser R D. Turbulence statistics in fully developed channel flow at low Reynolds number[J]. J Fluid Mech, 1987, 177: 133-166.
[7]	Nvidia. NVIDIA CUDA Programming Guide[K]. Version 2.0. 2008.
[8]	Ogawa S, Aoki T. GPU computing for 2dimensional incompressible-flow simulation based on multi-grid method[C]//Transactions of JSCES. Paper No20090021. 2009.
[9]	Harada T. Smoothed particle hydrodynamics on GPUs[C]//Proceeding of the Spring Conference on Computer Graphics, 2007: 235-241.
[10]	Rossinelli D, Bergdorf M, Cottet GH, Koumoutsakosa P. GPU accelerated simulations of bluff body flows using vortex particle methods[J]. Journal of Computational Physics, 2010, 229(9): 3316-3333.
[11]	Wang X, Aoki T. MultiGPU performance of incompressible flow computation by lattice Boltzmann method on GPU cluster[J]. Parallel Computing, 2011, 37(9): 521-535.
[12]	Shimokawabe T, Aoki T, Takaki T, Endo T, Yamanaka A, Maruyama N, Nukada A, Matsuoka S. Petascale phase-field simulation for dendritic solidification on the TSUBAME 2.0 supercomputer[C]//Proceedings of 2011 International Conference for High Performance Computing, Networking, Storage and Analysis. New York, USA, 2011.
[13]	Shimokawabe T, Aoki T, Ishida J, Kawano K, Muroi C. 145 TFlops performance on 3990 GPUs of TSUBAME 2.0 supercomputer for an operational weather prediction[C]// Proceedings of the International Conference on Computational Science, ICCS 〖STBX〗2011, 2011, 4: 1535-1544.
[14]	Wang X, Aoki T. High performance computation by multi-node GPU cluster-TSUBAME 2.0 on the air flow in an urban city using lattice Boltzmann method[J]. International Journal of Aerospace and Lightweight Structures, 2012, 2(1): 77-86.
[15]	Miki T, Wang X, Aoki T, Imai Y, Ishikawa T, Takase K, Yamaguchi T. Patientspecific modeling of pulmonary air flow using GPU cluster for the application in medical practice[J]. Computer Methods in Biomechanics and Biomedical Engineering, 2012, 15(7): 771-778.
[16]	Lammers P, Beronov K N, Volkert R, Brenner G, Durst F.Lattice BGK direct numerical simulation of fully developed turbulence in incompressible plane channel flow[J]. Computers & Fluids, 2006, 35(10): 1137-1153.
[17]	Jimenez J, Moin P. The minimal flow unit in near-wall turbulence[J]. Journal of Fluid Mechanics, 1991, 225(1): 213-240.
[18]	Spasov M, Rempfer D, Mokhasi P. Simulation of turbulent channel flow with an entropic lattice Boltzmann method[J]. Int J Numer Meth Fluids, 2009, 60(11): 1240-1258.
[19]	Pope S B. Turbulent Flows[M]. Cambridge: Cambridge University Press, 2000.

施引文献

资源附件(0)

访问统计

计量

文章访问数: 1796
HTML全文浏览量: 180
PDF下载量: 1392
被引次数: 0

姓名
邮箱
手机号码
标题
留言内容
验证码

留言板

基于多GPU的格子Boltzmann法对槽道湍流的直接数值模拟

doi: 10.3879/j.issn.1000-0887.2013.09.009

作者简介:
许丁 (1980—)，男，博士，讲师(E-mail:dingxu@mail.xjtu.edu.cn);王娴 (1977—)，女，博士，副教授（通讯作者.E-mail: wangxian@mail.xjtu.edu.cn）.

计量

Direct Numerical Simulation of the Wall-Bounded Turbulent Flow by Lattice Boltzmann Method Based on Multi-GPU

计量

目录

留言板

基于多GPU的格子Boltzmann法对槽道湍流的直接数值模拟

doi: 10.3879/j.issn.1000-0887.2013.09.009

作者简介: 许丁 (1980—)，男，博士，讲师(E-mail:dingxu@mail.xjtu.edu.cn);王娴 (1977—)，女，博士，副教授（通讯作者.E-mail: wangxian@mail.xjtu.edu.cn）.

计量

出版历程

Direct Numerical Simulation of the Wall-Bounded Turbulent Flow by Lattice Boltzmann Method Based on Multi-GPU

计量

出版历程

目录

作者简介:
许丁 (1980—)，男，博士，讲师(E-mail:dingxu@mail.xjtu.edu.cn);王娴 (1977—)，女，博士，副教授（通讯作者.E-mail: wangxian@mail.xjtu.edu.cn）.