Massively Parallel Finite Element Programming

Timo Heister; Martin Kronbichler; Wolfgang Bangerth

doi:10.1007/978-3-642-15646-5_13

Massively Parallel Finite Element Programming

Timo Heister, Martin Kronbichler, Wolfgang Bangerth

Research output: Chapter in Book/Report/Conference proceeding › Chapter

7 Scopus citations

Abstract

Today's large finite element simulations require parallel algorithms to scale on clusters with thousands or tens of thousands of processor cores. We present data structures and algorithms to take advantage of the power of high performance computers in generic finite element codes. Existing generic finite element libraries often restrict the parallelization to parallel linear algebra routines. This is a limiting factor when solving on more than a few hundreds of cores. We describe routines for distributed storage of all major components coupled with efficient, scalable algorithms. We give an overview of our effort to enable the modern and generic finite element library deal.II to take advantage of the power of large clusters. In particular, we describe the construction of a distributed mesh and develop algorithms to fully parallelize the finite element calculation. Numerical results demonstrate good scalability. © 2010 Springer-Verlag.

Original language	English (US)
Title of host publication	Recent Advances in the Message Passing Interface
Publisher	Springer Nature
Pages	122-131
Number of pages	10
ISBN (Print)	9783642156458
DOIs	https://doi.org/10.1007/978-3-642-15646-5_13
State	Published - 2010
Externally published	Yes

Access to Document

10.1007/978-3-642-15646-5_13

Cite this

@inbook{7f7fedce554c4ca7aa25b8e13ca264bf,

title = "Massively Parallel Finite Element Programming",

abstract = "Today's large finite element simulations require parallel algorithms to scale on clusters with thousands or tens of thousands of processor cores. We present data structures and algorithms to take advantage of the power of high performance computers in generic finite element codes. Existing generic finite element libraries often restrict the parallelization to parallel linear algebra routines. This is a limiting factor when solving on more than a few hundreds of cores. We describe routines for distributed storage of all major components coupled with efficient, scalable algorithms. We give an overview of our effort to enable the modern and generic finite element library deal.II to take advantage of the power of large clusters. In particular, we describe the construction of a distributed mesh and develop algorithms to fully parallelize the finite element calculation. Numerical results demonstrate good scalability. {\textcopyright} 2010 Springer-Verlag.",

author = "Timo Heister and Martin Kronbichler and Wolfgang Bangerth",

note = "KAUST Repository Item: Exported on 2020-10-01 Acknowledged KAUST grant number(s): KUS-C1-016-04 Acknowledgements: Timo Heister is partly supported by the German ResearchFoundation (DFG) through GK 1023. Martin Kronbichler is supported by theGraduate School in Mathematics and Computation (FMB). Wolfgang Bangerthwas partially supported by Award No. KUS-C1-016-04 made by King Abdullah University of Science and Technology (KAUST), by a grant from the NSF-funded Computational Infrastructure in Geodynamics initiative through AwardNo. EAR-0426271, and by an Alfred P. Sloan Research Fellowship.The computations were done on the Hurr cluster of the Institute for AppliedMathematics and Computational Science(IAMCS)atTexasA&MUniversity.Hurr is supported by Award No. KUS-C1-016-04 made by King Abdullah Uni-versity of Science and Technology (KAUST). This publication acknowledges KAUST support, but has no KAUST affiliated authors.",

year = "2010",

doi = "10.1007/978-3-642-15646-5_13",

language = "English (US)",

isbn = "9783642156458",

pages = "122--131",

booktitle = "Recent Advances in the Message Passing Interface",

publisher = "Springer Nature",

}

TY - CHAP

T1 - Massively Parallel Finite Element Programming

AU - Heister, Timo

AU - Kronbichler, Martin

AU - Bangerth, Wolfgang

N1 - KAUST Repository Item: Exported on 2020-10-01 Acknowledged KAUST grant number(s): KUS-C1-016-04 Acknowledgements: Timo Heister is partly supported by the German ResearchFoundation (DFG) through GK 1023. Martin Kronbichler is supported by theGraduate School in Mathematics and Computation (FMB). Wolfgang Bangerthwas partially supported by Award No. KUS-C1-016-04 made by King Abdullah University of Science and Technology (KAUST), by a grant from the NSF-funded Computational Infrastructure in Geodynamics initiative through AwardNo. EAR-0426271, and by an Alfred P. Sloan Research Fellowship.The computations were done on the Hurr cluster of the Institute for AppliedMathematics and Computational Science(IAMCS)atTexasA&MUniversity.Hurr is supported by Award No. KUS-C1-016-04 made by King Abdullah Uni-versity of Science and Technology (KAUST). This publication acknowledges KAUST support, but has no KAUST affiliated authors.

PY - 2010

Y1 - 2010

N2 - Today's large finite element simulations require parallel algorithms to scale on clusters with thousands or tens of thousands of processor cores. We present data structures and algorithms to take advantage of the power of high performance computers in generic finite element codes. Existing generic finite element libraries often restrict the parallelization to parallel linear algebra routines. This is a limiting factor when solving on more than a few hundreds of cores. We describe routines for distributed storage of all major components coupled with efficient, scalable algorithms. We give an overview of our effort to enable the modern and generic finite element library deal.II to take advantage of the power of large clusters. In particular, we describe the construction of a distributed mesh and develop algorithms to fully parallelize the finite element calculation. Numerical results demonstrate good scalability. © 2010 Springer-Verlag.

AB - Today's large finite element simulations require parallel algorithms to scale on clusters with thousands or tens of thousands of processor cores. We present data structures and algorithms to take advantage of the power of high performance computers in generic finite element codes. Existing generic finite element libraries often restrict the parallelization to parallel linear algebra routines. This is a limiting factor when solving on more than a few hundreds of cores. We describe routines for distributed storage of all major components coupled with efficient, scalable algorithms. We give an overview of our effort to enable the modern and generic finite element library deal.II to take advantage of the power of large clusters. In particular, we describe the construction of a distributed mesh and develop algorithms to fully parallelize the finite element calculation. Numerical results demonstrate good scalability. © 2010 Springer-Verlag.

UR - http://hdl.handle.net/10754/598761

UR - http://link.springer.com/10.1007/978-3-642-15646-5_13

UR - http://www.scopus.com/inward/record.url?scp=78149241169&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-15646-5_13

DO - 10.1007/978-3-642-15646-5_13

M3 - Chapter

SN - 9783642156458

SP - 122

EP - 131

BT - Recent Advances in the Message Passing Interface

PB - Springer Nature

ER -

Massively Parallel Finite Element Programming

Abstract

Access to Document

Other files and links

Fingerprint

Cite this