New multi-GPU implementation for smoothed particle hydrodynamics on heterogeneous clusters

被引：154
作者：
Dominguez, J. M. [1 ]
Crespo, A. J. C. [1 ]
Valdez-Balderas, D. [2 ]
Rogers, B. D. [2 ]
Gomez-Gesteira, M. [1 ]
机构：
[1] Univ Vigo, EPHYSLAB Environm Phys Lab, Vigo, Spain
[2] Univ Manchester, Sch Mech Aerosp & Civil Engn MACE, Modelling & Simulat Ctr MaSC, Manchester M13 9PL, Lancs, England
来源：
COMPUTER PHYSICS COMMUNICATIONS | 2013年 / 184卷 / 08期
基金：
英国工程与自然科学研究理事会;
关键词：
HPC; GPU; Multi-GPU; MPI; CUDA; SPH; Meshfree methods; SURFACE FLUID SOLVER; WAVE IMPACT; SIMULATION; SPHYSICS;
D O I：
10.1016/j.cpc.2013.03.008
中图分类号：
TP39 [计算机的应用];
学科分类号：
081203 ; 0835 ;
摘要：
A massively parallel SPH scheme using heterogeneous clusters of Central Processing Units (CPUs) and Graphics Processing Units (GPUs) has been developed. The new implementation originates from the single-GPU DualSPHysics code previously demonstrated to be powerful, stable and accurate. A combination of different parallel programming languages is combined to exploit not only one device (CPU or GPU) but also the combination of different machines. Communication among devices uses an improved Message Passing Interface (MPI) implementation which addresses some of the well-known drawbacks of MPI such as including a dynamic load balancing and overlapping data communications and computation tasks. The efficiency and scalability (strong and weak scaling) obtained with the new DualSPHysics code are analysed for different numbers of particles and different number of GPUs. Last, an application with more than 10(9) particles is presented to show the capability of the code to handle simulations that otherwise require large CPU clusters or supercomputers. (C) 2013 Elsevier B.V. All rights reserved.
引用
收藏
页码：1848 / 1860
页数：13
相关论文
共 22 条
[1]   Performance modeling of microsecond scale biological molecular dynamics simulations on heterogeneous architectures [J].
Agarwal, Pratul K. ;
Hampton, Scott ;
Poznanovic, Jeffrey ;
Ramanthan, Arvind ;
Alam, Sadaf R. ;
Crozier, Paul S. .
CONCURRENCY AND COMPUTATION-PRACTICE & EXPERIENCE, 2013, 25 (10) :1356-1375
[2]  
BATCHELOR GK, 1974, INTRO FLUID DYNAMICS
[3]   Implementing molecular dynamics on hybrid high performance computers - short range forces [J].
Brown, W. Michael ;
Wang, Peng ;
Plimpton, Steven J. ;
Tharrington, Arnold N. .
COMPUTER PHYSICS COMMUNICATIONS, 2011, 182 (04) :898-911
[4]   3D SPH Simulation of large waves mitigation with a dike [J].
Crespo, A. J. C. ;
Gomez-Gesteira, M. ;
Dalrymple, R. A. .
JOURNAL OF HYDRAULIC RESEARCH, 2007, 45 (05) :631-642
[5]   GPUs, a New Tool of Acceleration in CFD: Efficiency and Reliability on Smoothed Particle Hydrodynamics Methods [J].
Crespo, Alejandro C. ;
Dominguez, Jose M. ;
Barreiro, Anxo ;
Gomez-Gesteira, Moncho ;
Rogers, Benedict D. .
PLOS ONE, 2011, 6 (06)
[6]   Neighbour lists in smoothed particle hydrodynamics [J].
Dominguez, J. M. ;
Crespo, A. J. C. ;
Gomez-Gesteira, M. ;
Marongiu, J. C. .
INTERNATIONAL JOURNAL FOR NUMERICAL METHODS IN FLUIDS, 2011, 67 (12) :2026-2042
[7]   Optimization strategies for CPU and GPU implementations of a smoothed particle hydrodynamics method [J].
Dominguez, Jose M. ;
Crespo, Alejandro J. C. ;
Gomez-Gesteira, Moncho .
COMPUTER PHYSICS COMMUNICATIONS, 2013, 184 (03) :617-627
[8]  
Fleissner F, 2008, ADV PARALLEL COMPUT, V15, P37
[9]   Using a three-dimensional smoothed particle hydrodynamics method for wave impact on a tall structure [J].
Gómez-Gesteira, M ;
Dalrymple, RA .
JOURNAL OF WATERWAY PORT COASTAL AND OCEAN ENGINEERING, 2004, 130 (02) :63-69
[10]   SPHysics - development of a free-surface fluid solver - Part 1: Theory and formulations [J].
Gomez-Gesteira, M. ;
Rogers, B. D. ;
Crespo, A. J. C. ;
Dalrymple, R. A. ;
Narayanaswamy, M. ;
Dominguez, J. M. .
COMPUTERS & GEOSCIENCES, 2012, 48 :289-299
123