Journal Article

PreJuSER-17919

http://join2-wiki.gsi.de/foswiki/pub/Main/Artwork/join2_logo100x88.png A massively parallel, multi-disciplinary Barnes-Hut tree code for extreme-scale N-body simulations

Winkel, M.FZJ* ; Speck, R. ; Hübner, H.FZJ* ; Arnold, L.FZJ* ; Krause, R. ; Gibbon, P.FZJ*

2012
North Holland Publ. Co. Amsterdam

This record in other databases:  

Please use a persistent id in citations: doi:10.1016/j.cpc.2011.12.013

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 inch Supercomputing Centre and demonstrate scaling across 299,912 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. (C) 2011 Elsevier B.V. All rights reserved.

Keyword(s): J ; Parallel Barnes-Hut tree code (auto) ; Blue Gene/P (auto) ; Hybrid (auto) ; Load balancing (auto) ; Vortex methods (auto) ; Pthreads (auto)

Note: The authors gratefully acknowledge the helpful support by Julich Supercomputing Centre and the JSC staff, especially M. Stephan and J. Docter. This work was supported in part by the Alliance Program of the Helmholtz Association (HA216/EMMI), the BMBF project ScaFaCoS and the EU TEXT project, as well as additional computing time via the VSR project JZAM04. R.S. and R.K. would like to thank the Swiss Platform for High-Performance and High-Productivity Computing (HP2C) for funding and support.

---

Contributing Institute(s):

1. Jülich Supercomputing Centre (JSC)

Research Program(s):

1. Scientific Computing (FUEK411) (FUEK411)
2. 411 - Computational Science and Mathematical Methods (POF2-411) (POF2-411)

Appears in the scientific report 2012

---

Database coverage:
; Current Contents - Social and Behavioral Sciences ; JCR ; NCBI Molecular Biology Database ; Nationallizenz ; SCOPUS ; Science Citation Index ; Science Citation Index Expanded ; Thomson Reuters Master Journal List ; Web of Science Core Collection

---