帮助 关于我们

返回检索结果

耦合Nvidia/AMD两类GPU的格子玻尔兹曼模拟
Lattice Boltzmann simulation on Nvidia and AMD GPUs

查看参考文献14篇

文摘 利用图形处理单元(graphic processing unit,GPU)进行通用计算近年来得到关注,Nvidia和AMD公司已推出了各自的开发环境CUDA和ASC.很多计算在GPU上的速度远高于目前的CPU.格子玻尔兹曼方法(lattice Boltzmann method,LBM)作为一种网格上的粒子方法,对流动模拟具有良好的内在并行性,非常适合利用GPU进行大规模并行计算.本文提出了一种耦合Nvidia和AMD的两类GPU完成LBM凹槽流模拟的算法,对于两类GPU,在LBM的D2Q9模型下分别设计了相应的算法和程序,之后利用消息传递接口(message passing interface,MPI)协议通过多程序多数据流(multi-program multi-data,MPMD)模式使其能够联合计算,以充分发挥混合GPU集群系统的性能.通过GPU和CPU程序结果的比较,证实了GPU计算的正确性和所能带来的显著的加速比,为建设通用大规模GPU并行计算平台提供了重要参考.
其他语种文摘 General purpose computing on graphic processing units (GPUs) has received great attention recently. Both Nvidia and AMD have announced their own SDK, CUDA and ASC, respectively. Compared with CPUs, many applications can achieve high speed-up on GPUs. As a mesh-based particle method for flow simulation, lattice Boltzmann method (LBM) has perfect intrinsic parallelism that is very suitable for large-scale parallel computing. In this article, we put forward an algorithm for LB simulation of flow in grooved channel using the D2Q9 model, running on both Nvidia and AMD GPUs in the multi-program multi-data (MPMD) mode through message passing interface (MPI). The capability of hybrid GPU clusters can thus be fully utilized. The correctness and performance of the computation is analyzed through comparison with corresponding CPU implementation and significant speed-up of the GPU implementation has been demonstrated. The results provide valuable references to the establishment of GPU-based high performance computing (HPC) systems.
来源 科学通报 ,2009,54(20):3177-3184 【核心库】
关键词 GPGPU ; 格子波尔兹曼 ; Nvidia ; AMD ; 多程序多数据流 ; 联合计算
地址

中国科学院过程工程研究所, 多相复杂系统国家重点实验室, 北京, 100190

语种 中文
文献类型 研究性论文
ISSN 0023-074X
学科 自动化技术、计算机技术
基金 国家重大科研装备研制项目 ;  国家自然科学基金 ;  中国科学院知识创新工程项目
文献收藏号 CSCD:3885382

参考文献 共 14 共1页

1.  Nvidia. Nvidia CUDA Compute Unified Device Architecture Programming Guide,2007 被引 2    
2.  AMD. AMD Stream Computing-User Guide v1.1,2008 被引 1    
3.  Chen S. Lattice Boltzmann method for fluid flows. Annual Review of Fluid Mechanics,1998,30:329-364 被引 275    
4.  T(o)lke I. Implementation of a lattice Boltzmann kernel using the Compute Unified Device Architecture developed by Nvidia. Comput Visual Sci,2008 被引 1    
5.  Riegel E. Implementation of a Lattice-Boltzmann method for numerical fluid mechanics using the Nvidia CUDA technology. Comp Sci-Res Dev,2009:23 被引 1    
6.  Kaufman A. Implementing the lattice Boltzmann model on commodity graphics hardware. J Star Mech-Theor Exp,2009,6:P06016 被引 1    
7.  Succi S. The Lattice-Boltzmann Equation for Fluid Dynamics and Beyond,2001 被引 1    
8.  Harris S. An Introduction to the Theory of the Boltzmann Equation,2004 被引 1    
9.  Qian Y H. Lattice BGK models for Navier-Stokes equation. EUROPHYSICS LETTERS,1992,17:479-484 被引 387    
10.  多相复杂系统国家重点实验室多尺度离散模拟项目组. 基于GPU的多尺度离散模拟并行计算,2009 被引 7    
11.  Sukop M C. Lattice Boltzmann Modeling:An Introduction for Geoscientists and Engineers,2006 被引 11    
12.  AMD. Compute abstraction layer(CAL)technology.Intermediate Language(IL)-Reference Manual v2.0,2008 被引 1    
13.  Chen F. Multi-scale HPC system for multi-scale discrete simulation-Development and applicadon of a supercomputer with one petaflops peak performance in single precision. China Particuology,2009,7:332-335 被引 1    
14.  Kahan W. Further remarks on reducing truncation errors. COMMUNICATIONS OF THE ACM,1965,8:40 被引 7    
引证文献 6

1 李建江 一种单GPU程序向多GPU移植的模板化技术 计算机研究与发展,2010,47(12):2185-2191
被引 3

2 黄昌盛 基于CUDA的格子Boltzmann方法:算法设计与程序优化 科学通报,2011,56(28/29):2434-2444
被引 5

显示所有6篇文献

论文科学数据集
PlumX Metrics
相关文献

 作者相关
 关键词相关
 参考文献相关

iAuthor 链接
张云 0000-0002-5998-6500
版权所有 ©2008 中国科学院文献情报中心 制作维护:中国科学院文献情报中心
地址:北京中关村北四环西路33号 邮政编码:100190 联系电话:(010)82627496 E-mail:cscd@mail.las.ac.cn 京ICP备05002861号-4 | 京公网安备11010802043238号