Scaling Hypre's multigrid solvers to 100,000 cores

Allison H. Baker; Robert D. Falgout; Tzanio V. Kolev; Ulrike Meier Yang

doi:10.1007/978-1-4471-2437-5_13

Scaling Hypre's multigrid solvers to 100,000 cores

Allison H. Baker
, Robert D. Falgout
, Tzanio V. Kolev
, Ulrike Meier Yang

Application Scalability and Performance Group

Lawrence Livermore Natl. Laboratory

Research output: Chapter in Book/Report/Conference proceeding › Chapter › peer-review

110 Scopus citations

Abstract

The hypre software library (http://www.llnl.gov/CASC/hypre/) is a collection of high performance preconditioners and solvers for large sparse linear systems of equations on massively parallel machines. This paper investigates the scaling properties of several of the popular multigrid solvers and system building interfaces in hypre on two modern parallel platforms. We present scaling results on over 100,000 cores and even solve a problem with over a trillion unknowns.

Original language	English
Title of host publication	High-Performance Scientific Computing
Subtitle of host publication	Algorithms and Applications
Publisher	Springer-Verlag London Ltd
Pages	261-279
Number of pages	19
Volume	9781447124375
ISBN (Electronic)	9781447124375
ISBN (Print)	1447124367, 9781447124368
DOIs	https://doi.org/10.1007/978-1-4471-2437-5_13
State	Published - Oct 1 2012

Access to Document

10.1007/978-1-4471-2437-5_13

Cite this

@inbook{91f4725f33da4ff595aaaafa9af96c0f,

title = "Scaling Hypre's multigrid solvers to 100,000 cores",

abstract = "The hypre software library (http://www.llnl.gov/CASC/hypre/) is a collection of high performance preconditioners and solvers for large sparse linear systems of equations on massively parallel machines. This paper investigates the scaling properties of several of the popular multigrid solvers and system building interfaces in hypre on two modern parallel platforms. We present scaling results on over 100,000 cores and even solve a problem with over a trillion unknowns.",

author = "Baker, \{Allison H.\} and Falgout, \{Robert D.\} and Kolev, \{Tzanio V.\} and Yang, \{Ulrike Meier\}",

note = "Publisher Copyright: {\textcopyright} 2012 Springer-Verlag London Limited.",

year = "2012",

month = oct,

day = "1",

doi = "10.1007/978-1-4471-2437-5\_13",

language = "English",

isbn = "1447124367",

volume = "9781447124375",

pages = "261--279",

booktitle = "High-Performance Scientific Computing",

publisher = "Springer-Verlag London Ltd",

}

TY - CHAP

T1 - Scaling Hypre's multigrid solvers to 100,000 cores

AU - Baker, Allison H.

AU - Falgout, Robert D.

AU - Kolev, Tzanio V.

AU - Yang, Ulrike Meier

PY - 2012/10/1

Y1 - 2012/10/1

N2 - The hypre software library (http://www.llnl.gov/CASC/hypre/) is a collection of high performance preconditioners and solvers for large sparse linear systems of equations on massively parallel machines. This paper investigates the scaling properties of several of the popular multigrid solvers and system building interfaces in hypre on two modern parallel platforms. We present scaling results on over 100,000 cores and even solve a problem with over a trillion unknowns.

AB - The hypre software library (http://www.llnl.gov/CASC/hypre/) is a collection of high performance preconditioners and solvers for large sparse linear systems of equations on massively parallel machines. This paper investigates the scaling properties of several of the popular multigrid solvers and system building interfaces in hypre on two modern parallel platforms. We present scaling results on over 100,000 cores and even solve a problem with over a trillion unknowns.

UR - https://www.scopus.com/pages/publications/84949177527

U2 - 10.1007/978-1-4471-2437-5_13

DO - 10.1007/978-1-4471-2437-5_13

M3 - Chapter

AN - SCOPUS:84949177527

SN - 1447124367

SN - 9781447124368

VL - 9781447124375

SP - 261

EP - 279

BT - High-Performance Scientific Computing

PB - Springer-Verlag London Ltd

ER -

Scaling Hypre's multigrid solvers to 100,000 cores

Abstract

Access to Document

Other files and links

Fingerprint

Cite this