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

Abstract 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.

[1]  Peng Wang,et al.  Implementing molecular dynamics on hybrid high performance computers - short range forces , 2011, Comput. Phys. Commun..

[2]  Holger Wendland,et al.  Piecewise polynomial, positive definite and compactly supported radial functions of minimal degree , 1995, Adv. Comput. Math..

[3]  Moncho Gómez-Gesteira,et al.  Optimization strategies for CPU and GPU implementations of a smoothed particle hydrodynamics method , 2013, Comput. Phys. Commun..

[4]  G. Batchelor,et al.  An Introduction to Fluid Dynamics , 1968 .

[5]  David Le Touzé,et al.  SPH high-performance computing simulations of rigid solids impacting the free-surface of water , 2009 .

[6]  J. Monaghan.  Smoothed particle hydrodynamics , 2005 .

[7]  Sadaf R. Alam,et al.  Performance modeling of microsecond scale biological molecular dynamics simulations on heterogeneous architectures , 2013, Concurr. Comput. Pract. Exp..

[8]  P. Pedley,et al.  An Introduction to Fluid Dynamics , 1968 .

[9]  B. Rogers,et al.  GPUs, a New Tool of Acceleration in CFD: Efficiency and Reliability on Smoothed Particle Hydrodynamics Methods , 2011, PloS one.

[10]  L. Verlet.  Computer "Experiments" on Classical Fluids. I. Thermodynamical Properties of Lennard-Jones Molecules , 1967 .

[11]  Moncho Gómez-Gesteira,et al.  Neighbour lists in smoothed particle hydrodynamics , 2011 .

[12]  Tao Tang,et al.  Fast parallel cutoff pair interactions for molecular dynamics on heterogeneous systems , 2012 .

[13]  Robert A. Dalrymple,et al.  SPHysics - development of a free-surface fluid solver - Part 2: Efficiency and test cases , 2012, Comput. Geosci..

[14]  Damien Violeau,et al.  Fluid Mechanics and the SPH Method: Theory and Applications , 2012 .

[15]  Benedict D. Rogers,et al.  Towards accelerating smoothed particle hydrodynamics simulations for free-surface flows on multi-GPU clusters , 2012, J. Parallel Distributed Comput..

[16]  Robert A. Dalrymple,et al.  Using a Three-Dimensional Smoothed Particle Hydrodynamics Method for Wave Impact on a Tall Structure , 2004 .

[17]  Peter Eberhard,et al.  Load Balanced Parallel Simulation of Particle-Fluid DEM-SPH Systems with Moving Boundaries , 2007, PARCO.

[18]  Paul S. Crozier,et al.  General-purpose molecular dynamics simulations on GPU-based clusters , 2011 .

[19]  Robert A. Dalrymple,et al.  3D SPH Simulation of large waves mitigation with a dike , 2007 .

[20]  Arthur Veldman,et al.  A Volume-of-Fluid based simulation method for wave impact problems , 2005 .

[21]  Benedict D. Rogers,et al.  SPHysics - development of a free-surface fluid solver - Part 1: Theory and formulations , 2012, Comput. Geosci..

[22]  J. Monaghan,et al.  Solitary Waves on a Cretan Beach , 1999 .