Long range forces in a performance portable Molecular Dynamics framework

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

Molecular Dynamics (MD) codes predict the fundamental properties of matter by following the trajectories of a collection of interacting model particles. To exploit diverse modern manycore hardware, efficient codes must use all available parallelism. At the same time they need to be portable and easily extendible by the domain specialist (physicist/chemist) without detailed knowledge of this hardware. To address this challenge, we recently described a new Domain Specific Language (DSL) for the development of performance portable MD codes based on a "Separation of Concerns": a Python framework automatically generates efficient parallel code for a range of target architectures. Electrostatic interactions between charged particles are important in many physical systems and often dominate the runtime. Here we discuss the inclusion of long-range interaction algorithms in our code generation framework. These algorithms require global communications and careful consideration has to be given to any impact on parallel scalability. We implemented an Ewald summation algorithm for electrostatic forces, present scaling comparisons for different system sizes and compare to the performance of existing codes. We also report on further performance optimisations delivered with OpenMP shared memory parallelism.
Original languageEnglish
Title of host publicationParallel Computing is Everywhere
Subtitle of host publicationProceedings of the ParCo2017 Parallel Computing Conference
EditorsSanzio Bassini, Marco Danelutto, Patrizio Dazzi, Gerhard R. Joubert, Frans Peters
PublisherIOS Press
Pages37 - 46
Number of pages9
Volume32
ISBN (Electronic)978-1-61499-843-3
ISBN (Print)978-1-61499-842-6
DOIs
Publication statusPublished - Mar 2018
EventParCo 2017 Parallel Computing Conference - Bologna, Italy
Duration: 12 Sep 201715 Sep 2017

Publication series

NameAdvances in Parallel Computing
PublisherIOS Press
Volume32

Conference

ConferenceParCo 2017 Parallel Computing Conference
CountryItaly
CityBologna
Period12/09/1715/09/17

Fingerprint

Molecular dynamics
Hardware
Electrostatic force
Charged particles
Coulomb interactions
Scalability
Trajectories
Data storage equipment
Communication
Code generation

Keywords

  • Domain Specific Language
  • Electrostatic
  • Ewald Summation
  • Molecular Dynamics
  • Parallel Computing

ASJC Scopus subject areas

  • Computer Science(all)

Cite this

Saunders, W., Grant, J., & Müller, E. (2018). Long range forces in a performance portable Molecular Dynamics framework. In S. Bassini, M. Danelutto, P. Dazzi, G. R. Joubert, & F. Peters (Eds.), Parallel Computing is Everywhere: Proceedings of the ParCo2017 Parallel Computing Conference (Vol. 32, pp. 37 - 46). (Advances in Parallel Computing; Vol. 32). IOS Press. https://doi.org/10.3233/978-1-61499-843-3-37

Long range forces in a performance portable Molecular Dynamics framework. / Saunders, William; Grant, James; Müller, Eike.

Parallel Computing is Everywhere: Proceedings of the ParCo2017 Parallel Computing Conference. ed. / Sanzio Bassini; Marco Danelutto; Patrizio Dazzi; Gerhard R. Joubert; Frans Peters. Vol. 32 IOS Press, 2018. p. 37 - 46 (Advances in Parallel Computing; Vol. 32).

Research output: Chapter in Book/Report/Conference proceedingChapter

Saunders, W, Grant, J & Müller, E 2018, Long range forces in a performance portable Molecular Dynamics framework. in S Bassini, M Danelutto, P Dazzi, GR Joubert & F Peters (eds), Parallel Computing is Everywhere: Proceedings of the ParCo2017 Parallel Computing Conference. vol. 32, Advances in Parallel Computing, vol. 32, IOS Press, pp. 37 - 46, ParCo 2017 Parallel Computing Conference, Bologna, Italy, 12/09/17. https://doi.org/10.3233/978-1-61499-843-3-37
Saunders W, Grant J, Müller E. Long range forces in a performance portable Molecular Dynamics framework. In Bassini S, Danelutto M, Dazzi P, Joubert GR, Peters F, editors, Parallel Computing is Everywhere: Proceedings of the ParCo2017 Parallel Computing Conference. Vol. 32. IOS Press. 2018. p. 37 - 46. (Advances in Parallel Computing). https://doi.org/10.3233/978-1-61499-843-3-37
Saunders, William ; Grant, James ; Müller, Eike. / Long range forces in a performance portable Molecular Dynamics framework. Parallel Computing is Everywhere: Proceedings of the ParCo2017 Parallel Computing Conference. editor / Sanzio Bassini ; Marco Danelutto ; Patrizio Dazzi ; Gerhard R. Joubert ; Frans Peters. Vol. 32 IOS Press, 2018. pp. 37 - 46 (Advances in Parallel Computing).
@inbook{49d419624f3642ef837d14940337ef84,
title = "Long range forces in a performance portable Molecular Dynamics framework",
abstract = "Molecular Dynamics (MD) codes predict the fundamental properties of matter by following the trajectories of a collection of interacting model particles. To exploit diverse modern manycore hardware, efficient codes must use all available parallelism. At the same time they need to be portable and easily extendible by the domain specialist (physicist/chemist) without detailed knowledge of this hardware. To address this challenge, we recently described a new Domain Specific Language (DSL) for the development of performance portable MD codes based on a {"}Separation of Concerns{"}: a Python framework automatically generates efficient parallel code for a range of target architectures. Electrostatic interactions between charged particles are important in many physical systems and often dominate the runtime. Here we discuss the inclusion of long-range interaction algorithms in our code generation framework. These algorithms require global communications and careful consideration has to be given to any impact on parallel scalability. We implemented an Ewald summation algorithm for electrostatic forces, present scaling comparisons for different system sizes and compare to the performance of existing codes. We also report on further performance optimisations delivered with OpenMP shared memory parallelism.",
keywords = "Domain Specific Language, Electrostatic, Ewald Summation, Molecular Dynamics, Parallel Computing",
author = "William Saunders and James Grant and Eike M{\"u}ller",
year = "2018",
month = "3",
doi = "10.3233/978-1-61499-843-3-37",
language = "English",
isbn = "978-1-61499-842-6",
volume = "32",
series = "Advances in Parallel Computing",
publisher = "IOS Press",
pages = "37 -- 46",
editor = "Sanzio Bassini and Marco Danelutto and Patrizio Dazzi and Joubert, {Gerhard R.} and Frans Peters",
booktitle = "Parallel Computing is Everywhere",
address = "Netherlands",

}

TY - CHAP

T1 - Long range forces in a performance portable Molecular Dynamics framework

AU - Saunders, William

AU - Grant, James

AU - Müller, Eike

PY - 2018/3

Y1 - 2018/3

N2 - Molecular Dynamics (MD) codes predict the fundamental properties of matter by following the trajectories of a collection of interacting model particles. To exploit diverse modern manycore hardware, efficient codes must use all available parallelism. At the same time they need to be portable and easily extendible by the domain specialist (physicist/chemist) without detailed knowledge of this hardware. To address this challenge, we recently described a new Domain Specific Language (DSL) for the development of performance portable MD codes based on a "Separation of Concerns": a Python framework automatically generates efficient parallel code for a range of target architectures. Electrostatic interactions between charged particles are important in many physical systems and often dominate the runtime. Here we discuss the inclusion of long-range interaction algorithms in our code generation framework. These algorithms require global communications and careful consideration has to be given to any impact on parallel scalability. We implemented an Ewald summation algorithm for electrostatic forces, present scaling comparisons for different system sizes and compare to the performance of existing codes. We also report on further performance optimisations delivered with OpenMP shared memory parallelism.

AB - Molecular Dynamics (MD) codes predict the fundamental properties of matter by following the trajectories of a collection of interacting model particles. To exploit diverse modern manycore hardware, efficient codes must use all available parallelism. At the same time they need to be portable and easily extendible by the domain specialist (physicist/chemist) without detailed knowledge of this hardware. To address this challenge, we recently described a new Domain Specific Language (DSL) for the development of performance portable MD codes based on a "Separation of Concerns": a Python framework automatically generates efficient parallel code for a range of target architectures. Electrostatic interactions between charged particles are important in many physical systems and often dominate the runtime. Here we discuss the inclusion of long-range interaction algorithms in our code generation framework. These algorithms require global communications and careful consideration has to be given to any impact on parallel scalability. We implemented an Ewald summation algorithm for electrostatic forces, present scaling comparisons for different system sizes and compare to the performance of existing codes. We also report on further performance optimisations delivered with OpenMP shared memory parallelism.

KW - Domain Specific Language

KW - Electrostatic

KW - Ewald Summation

KW - Molecular Dynamics

KW - Parallel Computing

UR - http://www.scopus.com/inward/record.url?scp=85043595039&partnerID=8YFLogxK

U2 - 10.3233/978-1-61499-843-3-37

DO - 10.3233/978-1-61499-843-3-37

M3 - Chapter

SN - 978-1-61499-842-6

VL - 32

T3 - Advances in Parallel Computing

SP - 37

EP - 46

BT - Parallel Computing is Everywhere

A2 - Bassini, Sanzio

A2 - Danelutto, Marco

A2 - Dazzi, Patrizio

A2 - Joubert, Gerhard R.

A2 - Peters, Frans

PB - IOS Press

ER -