Performance analysis of the lattice Boltzmann model beyond Navier-Stokes

Amanda Peters Randles, Vivek Kale, Jeff Hammond, William Gropp, Efthimios Kaxiras

Research output: Contribution to conferencePaper

Abstract

The lattice Boltzmann method is increasingly important in facilitating large-scale fluid dynamics simulations. To date, these simulations have been built on discretized velocity models of up to 27 neighbors. Recent work has shown that higher order approximations of the continuum Boltzmann equation enable not only recovery of the Navier-Stokes hydro-dynamics, but also simulations for a wider range of Knudsen numbers, which is especially important in micro- and nano-scale flows. These higher-order models have significant impact on both the communication and computational complexity of the application. We present a performance study of the higher-order models as compared to the traditional ones, on both the IBM Blue Gene/P and Blue Gene/Q architectures. We study the tradeoffs of many optimizations methods such as the use of deep halo level ghost cells that, alongside hybrid programming models, reduce the impact of extended models and enable efficient modeling of extreme regimes of computational fluid dynamics.

Original languageEnglish (US)
Pages1063-1074
Number of pages12
DOIs
StatePublished - Oct 7 2013
Event27th IEEE International Parallel and Distributed Processing Symposium, IPDPS 2013 - Boston, MA, United States
Duration: May 20 2013May 24 2013

Other

Other27th IEEE International Parallel and Distributed Processing Symposium, IPDPS 2013
CountryUnited States
CityBoston, MA
Period5/20/135/24/13

Fingerprint

Genes
Boltzmann equation
Fluid dynamics
Computational complexity
Computational fluid dynamics
Recovery
Communication
Computer simulation

Keywords

  • fluid dynamics
  • lattice Boltzmann
  • multicore optimization

ASJC Scopus subject areas

  • Software

Cite this

Randles, A. P., Kale, V., Hammond, J., Gropp, W., & Kaxiras, E. (2013). Performance analysis of the lattice Boltzmann model beyond Navier-Stokes. 1063-1074. Paper presented at 27th IEEE International Parallel and Distributed Processing Symposium, IPDPS 2013, Boston, MA, United States. https://doi.org/10.1109/IPDPS.2013.109

Performance analysis of the lattice Boltzmann model beyond Navier-Stokes. / Randles, Amanda Peters; Kale, Vivek; Hammond, Jeff; Gropp, William; Kaxiras, Efthimios.

2013. 1063-1074 Paper presented at 27th IEEE International Parallel and Distributed Processing Symposium, IPDPS 2013, Boston, MA, United States.

Research output: Contribution to conferencePaper

Randles, AP, Kale, V, Hammond, J, Gropp, W & Kaxiras, E 2013, 'Performance analysis of the lattice Boltzmann model beyond Navier-Stokes', Paper presented at 27th IEEE International Parallel and Distributed Processing Symposium, IPDPS 2013, Boston, MA, United States, 5/20/13 - 5/24/13 pp. 1063-1074. https://doi.org/10.1109/IPDPS.2013.109
Randles AP, Kale V, Hammond J, Gropp W, Kaxiras E. Performance analysis of the lattice Boltzmann model beyond Navier-Stokes. 2013. Paper presented at 27th IEEE International Parallel and Distributed Processing Symposium, IPDPS 2013, Boston, MA, United States. https://doi.org/10.1109/IPDPS.2013.109
Randles, Amanda Peters ; Kale, Vivek ; Hammond, Jeff ; Gropp, William ; Kaxiras, Efthimios. / Performance analysis of the lattice Boltzmann model beyond Navier-Stokes. Paper presented at 27th IEEE International Parallel and Distributed Processing Symposium, IPDPS 2013, Boston, MA, United States.12 p.
@conference{fc2456448a5e4a51b5e8ba87e9591b1e,
title = "Performance analysis of the lattice Boltzmann model beyond Navier-Stokes",
abstract = "The lattice Boltzmann method is increasingly important in facilitating large-scale fluid dynamics simulations. To date, these simulations have been built on discretized velocity models of up to 27 neighbors. Recent work has shown that higher order approximations of the continuum Boltzmann equation enable not only recovery of the Navier-Stokes hydro-dynamics, but also simulations for a wider range of Knudsen numbers, which is especially important in micro- and nano-scale flows. These higher-order models have significant impact on both the communication and computational complexity of the application. We present a performance study of the higher-order models as compared to the traditional ones, on both the IBM Blue Gene/P and Blue Gene/Q architectures. We study the tradeoffs of many optimizations methods such as the use of deep halo level ghost cells that, alongside hybrid programming models, reduce the impact of extended models and enable efficient modeling of extreme regimes of computational fluid dynamics.",
keywords = "fluid dynamics, lattice Boltzmann, multicore optimization",
author = "Randles, {Amanda Peters} and Vivek Kale and Jeff Hammond and William Gropp and Efthimios Kaxiras",
year = "2013",
month = "10",
day = "7",
doi = "10.1109/IPDPS.2013.109",
language = "English (US)",
pages = "1063--1074",
note = "27th IEEE International Parallel and Distributed Processing Symposium, IPDPS 2013 ; Conference date: 20-05-2013 Through 24-05-2013",

}

TY - CONF

T1 - Performance analysis of the lattice Boltzmann model beyond Navier-Stokes

AU - Randles, Amanda Peters

AU - Kale, Vivek

AU - Hammond, Jeff

AU - Gropp, William

AU - Kaxiras, Efthimios

PY - 2013/10/7

Y1 - 2013/10/7

N2 - The lattice Boltzmann method is increasingly important in facilitating large-scale fluid dynamics simulations. To date, these simulations have been built on discretized velocity models of up to 27 neighbors. Recent work has shown that higher order approximations of the continuum Boltzmann equation enable not only recovery of the Navier-Stokes hydro-dynamics, but also simulations for a wider range of Knudsen numbers, which is especially important in micro- and nano-scale flows. These higher-order models have significant impact on both the communication and computational complexity of the application. We present a performance study of the higher-order models as compared to the traditional ones, on both the IBM Blue Gene/P and Blue Gene/Q architectures. We study the tradeoffs of many optimizations methods such as the use of deep halo level ghost cells that, alongside hybrid programming models, reduce the impact of extended models and enable efficient modeling of extreme regimes of computational fluid dynamics.

AB - The lattice Boltzmann method is increasingly important in facilitating large-scale fluid dynamics simulations. To date, these simulations have been built on discretized velocity models of up to 27 neighbors. Recent work has shown that higher order approximations of the continuum Boltzmann equation enable not only recovery of the Navier-Stokes hydro-dynamics, but also simulations for a wider range of Knudsen numbers, which is especially important in micro- and nano-scale flows. These higher-order models have significant impact on both the communication and computational complexity of the application. We present a performance study of the higher-order models as compared to the traditional ones, on both the IBM Blue Gene/P and Blue Gene/Q architectures. We study the tradeoffs of many optimizations methods such as the use of deep halo level ghost cells that, alongside hybrid programming models, reduce the impact of extended models and enable efficient modeling of extreme regimes of computational fluid dynamics.

KW - fluid dynamics

KW - lattice Boltzmann

KW - multicore optimization

UR - http://www.scopus.com/inward/record.url?scp=84884861013&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84884861013&partnerID=8YFLogxK

U2 - 10.1109/IPDPS.2013.109

DO - 10.1109/IPDPS.2013.109

M3 - Paper

AN - SCOPUS:84884861013

SP - 1063

EP - 1074

ER -