A massively parallel geometric multigrid solver on hierarchically distributed grids

Sebastian Reiter; Andreas Vogel; Ingo Heppner; Martin Rupp; Gabriel Wittum

doi:10.1007/s00791-014-0231-x

A massively parallel geometric multigrid solver on hierarchically distributed grids

Sebastian Reiter^*, Andreas Vogel, Ingo Heppner, Martin Rupp, Gabriel Wittum

^*Corresponding author for this work

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

56 Scopus citations

Abstract

A parallel geometric multigrid solver on hierarchically distributed grids is presented. Using a tree-structure for grid distribution onto the processing entities, the multigrid cycle is performed similarly to the serial algorithm, using additional vertical communication during transfer operations. The workload is gathered to fewer processes on coarser levels. Involved parallel structures are described in detail and the multigrid algorithm is formulated, discussing parallelization details. A performance study is presented that shows close to optimal efficiency for weak scaling up to 262k processes in 2 and 3 space dimensions.

Original language	English (US)
Pages (from-to)	151-164
Number of pages	14
Journal	Computing and Visualization in Science
Volume	16
Issue number	4
DOIs	https://doi.org/10.1007/s00791-014-0231-x
State	Published - Aug 2013
Externally published	Yes

Bibliographical note

Publisher Copyright:
© 2014, Springer-Verlag Berlin Heidelberg.

Keywords

Multigrid methods
Parallel geometric multigrid
Parallelization
Weak scaling

ASJC Scopus subject areas

Theoretical Computer Science
Software
Modeling and Simulation
General Engineering
Computer Vision and Pattern Recognition
Computational Theory and Mathematics

Access to Document

10.1007/s00791-014-0231-x

Cite this

@article{0208d71334cb4be1b42f41d6a2410aa5,

title = "A massively parallel geometric multigrid solver on hierarchically distributed grids",

abstract = "A parallel geometric multigrid solver on hierarchically distributed grids is presented. Using a tree-structure for grid distribution onto the processing entities, the multigrid cycle is performed similarly to the serial algorithm, using additional vertical communication during transfer operations. The workload is gathered to fewer processes on coarser levels. Involved parallel structures are described in detail and the multigrid algorithm is formulated, discussing parallelization details. A performance study is presented that shows close to optimal efficiency for weak scaling up to 262k processes in 2 and 3 space dimensions.",

keywords = "Multigrid methods, Parallel geometric multigrid, Parallelization, Weak scaling",

author = "Sebastian Reiter and Andreas Vogel and Ingo Heppner and Martin Rupp and Gabriel Wittum",

note = "Publisher Copyright: {\textcopyright} 2014, Springer-Verlag Berlin Heidelberg.",

year = "2013",

month = aug,

doi = "10.1007/s00791-014-0231-x",

language = "English (US)",

volume = "16",

pages = "151--164",

journal = "Computing and Visualization in Science",

issn = "1432-9360",

publisher = "Springer Verlag",

number = "4",

}

TY - JOUR

T1 - A massively parallel geometric multigrid solver on hierarchically distributed grids

AU - Reiter, Sebastian

AU - Vogel, Andreas

AU - Heppner, Ingo

AU - Rupp, Martin

AU - Wittum, Gabriel

PY - 2013/8

Y1 - 2013/8

N2 - A parallel geometric multigrid solver on hierarchically distributed grids is presented. Using a tree-structure for grid distribution onto the processing entities, the multigrid cycle is performed similarly to the serial algorithm, using additional vertical communication during transfer operations. The workload is gathered to fewer processes on coarser levels. Involved parallel structures are described in detail and the multigrid algorithm is formulated, discussing parallelization details. A performance study is presented that shows close to optimal efficiency for weak scaling up to 262k processes in 2 and 3 space dimensions.

AB - A parallel geometric multigrid solver on hierarchically distributed grids is presented. Using a tree-structure for grid distribution onto the processing entities, the multigrid cycle is performed similarly to the serial algorithm, using additional vertical communication during transfer operations. The workload is gathered to fewer processes on coarser levels. Involved parallel structures are described in detail and the multigrid algorithm is formulated, discussing parallelization details. A performance study is presented that shows close to optimal efficiency for weak scaling up to 262k processes in 2 and 3 space dimensions.

KW - Multigrid methods

KW - Parallel geometric multigrid

KW - Parallelization

KW - Weak scaling

UR - http://www.scopus.com/inward/record.url?scp=84925466854&partnerID=8YFLogxK

U2 - 10.1007/s00791-014-0231-x

DO - 10.1007/s00791-014-0231-x

M3 - Article

AN - SCOPUS:84925466854

SN - 1432-9360

VL - 16

SP - 151

EP - 164

JO - Computing and Visualization in Science

JF - Computing and Visualization in Science

IS - 4

ER -

A massively parallel geometric multigrid solver on hierarchically distributed grids

Abstract

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this