Massively parallel simulations for disordered systems

Ravinder Kumar; Jonathan Gross; Wolfhard Janke; Martin Weigel

doi:10.1140/epjb/e2020-100535-0

Massively parallel simulations for disordered systems

Ravinder Kumar, Jonathan Gross, Wolfhard Janke, Martin Weigel

School of Computing, Electronics and Maths

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

1 Citation (Scopus)

4 Downloads (Pure)

Abstract

Abstract: Simulations of systems with quenched disorder are extremely demanding, suffering from the combined effect of slow relaxation and the need of performing the disorder average. As a consequence, new algorithms and improved implementations in combination with alternative and even purpose-built hardware are often instrumental for conducting meaningful studies of such systems. The ensuing demands regarding hardware availability and code complexity are substantial and sometimes prohibitive. We demonstrate how with a moderate coding effort leaving the overall structure of the simulation code unaltered as compared to a CPU implementation, very significant speed-ups can be achieved from a parallel code on GPU by mainly exploiting the trivial parallelism of the disorder samples and the near-trivial parallelism of the parallel tempering replicas. A combination of this massively parallel implementation with a careful choice of the temperature protocol for parallel tempering as well as efficient cluster updates allows us to equilibrate comparatively large systems with moderate computational resources. Graphical abstract: [Figure not available: see fulltext.].

Original language	English
Article number	79
Journal	The European Physical Journal B - Condensed Matter and Complex Systems
Volume	93
Issue number	5
Early online date	4 May 2020
DOIs	https://doi.org/10.1140/epjb/e2020-100535-0
Publication status	Published - May 2020

Bibliographical note

The final publication is available at Springer via http://dx.doi.org/10.1140/epjb/e2020-100535-0

Copyright © and Moral Rights are retained by the author(s) and/ or other copyright owners. A copy can be downloaded for personal non-commercial research or study, without prior permission or charge. This item cannot be reproduced or quoted extensively from without first obtaining permission in writing from the copyright holder(s). The content must not be changed in any way or sold commercially in any format or medium without the formal permission of the copyright holders.

Keywords

physics.comp-ph
cond-mat.dis-nn
cond-mat.stat-mech

Access to Document

10.1140/epjb/e2020-100535-0

Post-PrintAccepted author manuscript, 525 KB

Cite this

@article{721d8515f41e4ac8b31091a2cbfa45ab,

title = "Massively parallel simulations for disordered systems",

abstract = "Abstract: Simulations of systems with quenched disorder are extremely demanding, suffering from the combined effect of slow relaxation and the need of performing the disorder average. As a consequence, new algorithms and improved implementations in combination with alternative and even purpose-built hardware are often instrumental for conducting meaningful studies of such systems. The ensuing demands regarding hardware availability and code complexity are substantial and sometimes prohibitive. We demonstrate how with a moderate coding effort leaving the overall structure of the simulation code unaltered as compared to a CPU implementation, very significant speed-ups can be achieved from a parallel code on GPU by mainly exploiting the trivial parallelism of the disorder samples and the near-trivial parallelism of the parallel tempering replicas. A combination of this massively parallel implementation with a careful choice of the temperature protocol for parallel tempering as well as efficient cluster updates allows us to equilibrate comparatively large systems with moderate computational resources. Graphical abstract: [Figure not available: see fulltext.].",

keywords = "physics.comp-ph, cond-mat.dis-nn, cond-mat.stat-mech",

author = "Ravinder Kumar and Jonathan Gross and Wolfhard Janke and Martin Weigel",

note = "The final publication is available at Springer via http://dx.doi.org/10.1140/epjb/e2020-100535-0 Copyright {\textcopyright} and Moral Rights are retained by the author(s) and/ or other copyright owners. A copy can be downloaded for personal non-commercial research or study, without prior permission or charge. This item cannot be reproduced or quoted extensively from without first obtaining permission in writing from the copyright holder(s). The content must not be changed in any way or sold commercially in any format or medium without the formal permission of the copyright holders. ",

year = "2020",

month = may,

doi = "10.1140/epjb/e2020-100535-0",

language = "English",

volume = "93",

journal = "The European Physical Journal B - Condensed Matter and Complex Systems",

issn = "1434-6028",

publisher = "EDP Sciences",

number = "5",

}

TY - JOUR

T1 - Massively parallel simulations for disordered systems

AU - Kumar, Ravinder

AU - Gross, Jonathan

AU - Janke, Wolfhard

AU - Weigel, Martin

N1 - The final publication is available at Springer via http://dx.doi.org/10.1140/epjb/e2020-100535-0 Copyright © and Moral Rights are retained by the author(s) and/ or other copyright owners. A copy can be downloaded for personal non-commercial research or study, without prior permission or charge. This item cannot be reproduced or quoted extensively from without first obtaining permission in writing from the copyright holder(s). The content must not be changed in any way or sold commercially in any format or medium without the formal permission of the copyright holders.

PY - 2020/5

Y1 - 2020/5

N2 - Abstract: Simulations of systems with quenched disorder are extremely demanding, suffering from the combined effect of slow relaxation and the need of performing the disorder average. As a consequence, new algorithms and improved implementations in combination with alternative and even purpose-built hardware are often instrumental for conducting meaningful studies of such systems. The ensuing demands regarding hardware availability and code complexity are substantial and sometimes prohibitive. We demonstrate how with a moderate coding effort leaving the overall structure of the simulation code unaltered as compared to a CPU implementation, very significant speed-ups can be achieved from a parallel code on GPU by mainly exploiting the trivial parallelism of the disorder samples and the near-trivial parallelism of the parallel tempering replicas. A combination of this massively parallel implementation with a careful choice of the temperature protocol for parallel tempering as well as efficient cluster updates allows us to equilibrate comparatively large systems with moderate computational resources. Graphical abstract: [Figure not available: see fulltext.].

AB - Abstract: Simulations of systems with quenched disorder are extremely demanding, suffering from the combined effect of slow relaxation and the need of performing the disorder average. As a consequence, new algorithms and improved implementations in combination with alternative and even purpose-built hardware are often instrumental for conducting meaningful studies of such systems. The ensuing demands regarding hardware availability and code complexity are substantial and sometimes prohibitive. We demonstrate how with a moderate coding effort leaving the overall structure of the simulation code unaltered as compared to a CPU implementation, very significant speed-ups can be achieved from a parallel code on GPU by mainly exploiting the trivial parallelism of the disorder samples and the near-trivial parallelism of the parallel tempering replicas. A combination of this massively parallel implementation with a careful choice of the temperature protocol for parallel tempering as well as efficient cluster updates allows us to equilibrate comparatively large systems with moderate computational resources. Graphical abstract: [Figure not available: see fulltext.].

KW - physics.comp-ph

KW - cond-mat.dis-nn

KW - cond-mat.stat-mech

UR - http://www.scopus.com/inward/record.url?scp=85084113441&partnerID=8YFLogxK

U2 - 10.1140/epjb/e2020-100535-0

DO - 10.1140/epjb/e2020-100535-0

M3 - Article

VL - 93

JO - The European Physical Journal B - Condensed Matter and Complex Systems

JF - The European Physical Journal B - Condensed Matter and Complex Systems

SN - 1434-6028

IS - 5

M1 - 79

ER -

Massively parallel simulations for disordered systems

Abstract

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this