A massively parallel, multi-disciplinary Barnes-Hut tree code for extreme-scale N-body simulations

Mathias Winkel; Robert Speck; Helge Hübner; Lukas Arnold; Rolf Krause; Paul Gibbon

doi:10.1016/j.cpc.2011.12.013

A massively parallel, multi-disciplinary Barnes-Hut tree code for extreme-scale N-body simulations

Mathias Winkel^*, Robert Speck, Helge Hübner, Lukas Arnold, Rolf Krause, Paul Gibbon

^*Corresponding author for this work

Applied Mathematics and Computational Science

Research output: Contribution to journal › Article › peer-review

60 Scopus citations

Abstract

The efficient parallelization of fast multipole-based algorithms for the N-body problem is one of the most challenging topics in high performance scientific computing. The emergence of non-local, irregular communication patterns generated by these algorithms can easily create an insurmountable bottleneck on supercomputers with hundreds of thousands of cores. To overcome this obstacle we have developed an innovative parallelization strategy for Barnes-Hut tree codes on present and upcoming HPC multicore architectures. This scheme, based on a combined MPI-Pthreads approach, permits an efficient overlap of computation and data exchange. We highlight the capabilities of this method on the full IBM Blue Gene/P system JUGENE at Jülich Supercomputing Centre and demonstrate scaling across 294,912 cores with up to 2,048,000,000 particles. Applying our implementation pepc to laser-plasma interaction and vortex particle methods close to the continuum limit, we demonstrate its potential for ground-breaking advances in large-scale particle simulations.

Original language	English (US)
Pages (from-to)	880-889
Number of pages	10
Journal	Computer Physics Communications
Volume	183
Issue number	4
DOIs	https://doi.org/10.1016/j.cpc.2011.12.013
State	Published - Apr 2012

Keywords

Blue Gene/P
Hybrid
Load balancing
Parallel Barnes-Hut tree code
Pthreads
Vortex methods

ASJC Scopus subject areas

Hardware and Architecture
General Physics and Astronomy

Access to Document

10.1016/j.cpc.2011.12.013

Cite this

@article{0c00f58cefd24477ba75352456ac0607,

title = "A massively parallel, multi-disciplinary Barnes-Hut tree code for extreme-scale N-body simulations",

abstract = "The efficient parallelization of fast multipole-based algorithms for the N-body problem is one of the most challenging topics in high performance scientific computing. The emergence of non-local, irregular communication patterns generated by these algorithms can easily create an insurmountable bottleneck on supercomputers with hundreds of thousands of cores. To overcome this obstacle we have developed an innovative parallelization strategy for Barnes-Hut tree codes on present and upcoming HPC multicore architectures. This scheme, based on a combined MPI-Pthreads approach, permits an efficient overlap of computation and data exchange. We highlight the capabilities of this method on the full IBM Blue Gene/P system JUGENE at J{\"u}lich Supercomputing Centre and demonstrate scaling across 294,912 cores with up to 2,048,000,000 particles. Applying our implementation pepc to laser-plasma interaction and vortex particle methods close to the continuum limit, we demonstrate its potential for ground-breaking advances in large-scale particle simulations.",

keywords = "Blue Gene/P, Hybrid, Load balancing, Parallel Barnes-Hut tree code, Pthreads, Vortex methods",

author = "Mathias Winkel and Robert Speck and Helge H{\"u}bner and Lukas Arnold and Rolf Krause and Paul Gibbon",

year = "2012",

month = apr,

doi = "10.1016/j.cpc.2011.12.013",

language = "English (US)",

volume = "183",

pages = "880--889",

journal = "Computer Physics Communications",

issn = "0010-4655",

publisher = "Elsevier B.V.",

number = "4",

}

TY - JOUR

T1 - A massively parallel, multi-disciplinary Barnes-Hut tree code for extreme-scale N-body simulations

AU - Winkel, Mathias

AU - Speck, Robert

AU - Hübner, Helge

AU - Arnold, Lukas

AU - Krause, Rolf

AU - Gibbon, Paul

PY - 2012/4

Y1 - 2012/4

N2 - The efficient parallelization of fast multipole-based algorithms for the N-body problem is one of the most challenging topics in high performance scientific computing. The emergence of non-local, irregular communication patterns generated by these algorithms can easily create an insurmountable bottleneck on supercomputers with hundreds of thousands of cores. To overcome this obstacle we have developed an innovative parallelization strategy for Barnes-Hut tree codes on present and upcoming HPC multicore architectures. This scheme, based on a combined MPI-Pthreads approach, permits an efficient overlap of computation and data exchange. We highlight the capabilities of this method on the full IBM Blue Gene/P system JUGENE at Jülich Supercomputing Centre and demonstrate scaling across 294,912 cores with up to 2,048,000,000 particles. Applying our implementation pepc to laser-plasma interaction and vortex particle methods close to the continuum limit, we demonstrate its potential for ground-breaking advances in large-scale particle simulations.

AB - The efficient parallelization of fast multipole-based algorithms for the N-body problem is one of the most challenging topics in high performance scientific computing. The emergence of non-local, irregular communication patterns generated by these algorithms can easily create an insurmountable bottleneck on supercomputers with hundreds of thousands of cores. To overcome this obstacle we have developed an innovative parallelization strategy for Barnes-Hut tree codes on present and upcoming HPC multicore architectures. This scheme, based on a combined MPI-Pthreads approach, permits an efficient overlap of computation and data exchange. We highlight the capabilities of this method on the full IBM Blue Gene/P system JUGENE at Jülich Supercomputing Centre and demonstrate scaling across 294,912 cores with up to 2,048,000,000 particles. Applying our implementation pepc to laser-plasma interaction and vortex particle methods close to the continuum limit, we demonstrate its potential for ground-breaking advances in large-scale particle simulations.

KW - Blue Gene/P

KW - Hybrid

KW - Load balancing

KW - Parallel Barnes-Hut tree code

KW - Pthreads

KW - Vortex methods

UR - http://www.scopus.com/inward/record.url?scp=84856228982&partnerID=8YFLogxK

U2 - 10.1016/j.cpc.2011.12.013

DO - 10.1016/j.cpc.2011.12.013

M3 - Article

AN - SCOPUS:84856228982

SN - 0010-4655

VL - 183

SP - 880

EP - 889

JO - Computer Physics Communications

JF - Computer Physics Communications

IS - 4

ER -

A massively parallel, multi-disciplinary Barnes-Hut tree code for extreme-scale N-body simulations

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this