A performance analysis of ensemble averaging for high fidelity turbulence simulations at the strong scaling limit

Vakhtang Makarashvili; Elia Merzari; Aleksandr Obabko; Andrew Siegel; Paul Fischer

doi:10.1016/j.cpc.2017.05.023

A performance analysis of ensemble averaging for high fidelity turbulence simulations at the strong scaling limit

Vakhtang Makarashvili, Elia Merzari, Aleksandr Obabko, Andrew Siegel, Paul Fischer

Ken and Mary Alice Lindquist Department of Nuclear Engineering

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

10 Scopus citations

Abstract

We analyze the potential performance benefits of estimating expected quantities in large eddy simulations of turbulent flows using true ensembles rather than ergodic time averaging. Multiple realizations of the same flow are simulated in parallel, using slightly perturbed initial conditions to create unique instantaneous evolutions of the flow field. Each realization is then used to calculate statistical quantities. Provided each instance is sufficiently de-correlated, this approach potentially allows considerable reduction in the time to solution beyond the strong scaling limit for a given accuracy. This paper focuses on the theory and implementation of the methodology in Nek5000, a massively parallel open-source spectral element code.

Original language	English (US)
Pages (from-to)	236-245
Number of pages	10
Journal	Computer Physics Communications
Volume	219
DOIs	https://doi.org/10.1016/j.cpc.2017.05.023
State	Published - Oct 2017

All Science Journal Classification (ASJC) codes

Hardware and Architecture
Physics and Astronomy(all)

Access to Document

10.1016/j.cpc.2017.05.023

Cite this

@article{461d4150d0614797b8703660b8f6f11f,

title = "A performance analysis of ensemble averaging for high fidelity turbulence simulations at the strong scaling limit",

abstract = "We analyze the potential performance benefits of estimating expected quantities in large eddy simulations of turbulent flows using true ensembles rather than ergodic time averaging. Multiple realizations of the same flow are simulated in parallel, using slightly perturbed initial conditions to create unique instantaneous evolutions of the flow field. Each realization is then used to calculate statistical quantities. Provided each instance is sufficiently de-correlated, this approach potentially allows considerable reduction in the time to solution beyond the strong scaling limit for a given accuracy. This paper focuses on the theory and implementation of the methodology in Nek5000, a massively parallel open-source spectral element code.",

author = "Vakhtang Makarashvili and Elia Merzari and Aleksandr Obabko and Andrew Siegel and Paul Fischer",

note = "Funding Information: This material was based upon work supported by the U.S. Department of Energy, Office of Science, as part of the CESAR Program, under contract No. DE-AC02-06CH11357. Publisher Copyright: {\textcopyright} 2017 Elsevier B.V.",

year = "2017",

month = oct,

doi = "10.1016/j.cpc.2017.05.023",

language = "English (US)",

volume = "219",

pages = "236--245",

journal = "Computer Physics Communications",

issn = "0010-4655",

publisher = "Elsevier",

}

TY - JOUR

T1 - A performance analysis of ensemble averaging for high fidelity turbulence simulations at the strong scaling limit

AU - Makarashvili, Vakhtang

AU - Merzari, Elia

AU - Obabko, Aleksandr

AU - Siegel, Andrew

AU - Fischer, Paul

N1 - Funding Information: This material was based upon work supported by the U.S. Department of Energy, Office of Science, as part of the CESAR Program, under contract No. DE-AC02-06CH11357. Publisher Copyright: © 2017 Elsevier B.V.

PY - 2017/10

Y1 - 2017/10

N2 - We analyze the potential performance benefits of estimating expected quantities in large eddy simulations of turbulent flows using true ensembles rather than ergodic time averaging. Multiple realizations of the same flow are simulated in parallel, using slightly perturbed initial conditions to create unique instantaneous evolutions of the flow field. Each realization is then used to calculate statistical quantities. Provided each instance is sufficiently de-correlated, this approach potentially allows considerable reduction in the time to solution beyond the strong scaling limit for a given accuracy. This paper focuses on the theory and implementation of the methodology in Nek5000, a massively parallel open-source spectral element code.

AB - We analyze the potential performance benefits of estimating expected quantities in large eddy simulations of turbulent flows using true ensembles rather than ergodic time averaging. Multiple realizations of the same flow are simulated in parallel, using slightly perturbed initial conditions to create unique instantaneous evolutions of the flow field. Each realization is then used to calculate statistical quantities. Provided each instance is sufficiently de-correlated, this approach potentially allows considerable reduction in the time to solution beyond the strong scaling limit for a given accuracy. This paper focuses on the theory and implementation of the methodology in Nek5000, a massively parallel open-source spectral element code.

UR - http://www.scopus.com/inward/record.url?scp=85023771083&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85023771083&partnerID=8YFLogxK

U2 - 10.1016/j.cpc.2017.05.023

DO - 10.1016/j.cpc.2017.05.023

M3 - Article

AN - SCOPUS:85023771083

SN - 0010-4655

VL - 219

SP - 236

EP - 245

JO - Computer Physics Communications

JF - Computer Physics Communications

ER -

A performance analysis of ensemble averaging for high fidelity turbulence simulations at the strong scaling limit

Abstract

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this