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

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

doi:10.1016/j.cpc.2017.05.023

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

Full Record
Other Related Research

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 study focuses on the theory and implementation of the methodology in Nek5000, a massively parallel open-source spectral element code.

Authors:

Makarashvili, Vakhtang ^[1]; Merzari, Elia ^[1]; Obabko, Aleksandr ^[1]; Siegel, Andrew ^[1]; Fischer, Paul ^[1]

Argonne National Lab. (ANL), Lemont, IL (United States)

Publication Date:: 2017-06-07

Research Org.:: Argonne National Lab. (ANL), Argonne, IL (United States)

Sponsoring Org.:: USDOE Office of Science (SC)

OSTI Identifier:: 1393517

Alternate Identifier(s):: OSTI ID: 1550297

Grant/Contract Number:: AC02-06CH11357

Resource Type:: Accepted Manuscript

Journal Name:: Computer Physics Communications

Additional Journal Information:: Journal Volume: 219; Journal Issue: C; Journal ID: ISSN 0010-4655

Publisher:: Elsevier

Country of Publication:: United States

Language:: English

Subject:: 71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSICS; 97 MATHEMATICS AND COMPUTING; Ensemble averaging; Nek5000

Citation Formats


                    Makarashvili, Vakhtang, Merzari, Elia, Obabko, Aleksandr, Siegel, Andrew, and Fischer, Paul. A performance analysis of ensemble averaging for high fidelity turbulence simulations at the strong scaling limit.  United States: N. p., 2017. 
Web.  doi:10.1016/j.cpc.2017.05.023.

Copy to clipboard


                    Makarashvili, Vakhtang, Merzari, Elia, Obabko, Aleksandr, Siegel, Andrew, & Fischer, Paul. A performance analysis of ensemble averaging for high fidelity turbulence simulations at the strong scaling limit.  United States.  https://doi.org/10.1016/j.cpc.2017.05.023

Copy to clipboard


                    Makarashvili, Vakhtang, Merzari, Elia, Obabko, Aleksandr, Siegel, Andrew, and Fischer, Paul. Wed .  
"A performance analysis of ensemble averaging for high fidelity turbulence simulations at the strong scaling limit".  United States.  https://doi.org/10.1016/j.cpc.2017.05.023.  https://www.osti.gov/servlets/purl/1393517.

Copy to clipboard


                    
@article{osti_1393517,

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

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

  abstractNote = {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 study focuses on the theory and implementation of the methodology in Nek5000, a massively parallel open-source spectral element code.},

  doi          = {10.1016/j.cpc.2017.05.023},

  journal      = {Computer Physics Communications},

  number       = C,

  volume       = 219,

  place        = {United States},

  year         = {2017},

  month        = {6}

}

Copy to clipboard

Journal Article:

Free Publicly Available Full Text

Accepted Manuscript (Publisher)

Accepted Manuscript (DOE)

Publisher's Version of Record

https://doi.org/10.1016/j.cpc.2017.05.023

Other availability

Search WorldCat to find libraries that may hold this journal

Citation Metrics:

Cited by: 7 works

Citation information provided by
Web of Science

Save / Share:

Export Metadata

Save to My Library

Similar Records in DOE PAGES and OSTI.GOV collections:

Filtering, averaging, and scale dependency in homogeneous variable density turbulence

Journal Article Saenz, Juan Antonio ; Aslangil, Denis ; Livescu, Daniel - Physics of Fluids

We invesmore »« less
https://doi.org/10.1063/5.0040337

Full Text Available
A high-fidelity method to analyze perturbation evolution in turbulent flows

Journal Article Unnikrishnan, S., E-mail: sasidharannair.1@osu.edu ; Gaitonde, Datta V., E-mail: gaitonde.3@osu.edu - Journal of Computational Physics

Small perturbation propagation in fluid flows is usually examined by linearizing the governing equations about a steady basic state. It is often useful, however, to study perturbation evolution in the unsteady evolving turbulent environment. Such analyses can elucidate the role of perturbations in the generation of coherent structures or the production of noise from jet turbulence. The appropriate equations are still the linearized Navier–Stokes equations, except that the linearization must be performed about the instantaneous evolving turbulent state, which forms the coefficients of the linearized equations. This is a far more difficult problem since in addition to the turbulent state,more »« less
https://doi.org/10.1016/J.JCP.2016.01.017
Kinetic theory of particle-in-cell simulation plasma and the ensemble averaging technique

Journal Article Touati, Michaël ; Codur, Romain ; Tsung, Frank ; ... - Plasma Physics and Controlled Fusion

Abstract We derive the kinetic theory of fluctuations in physically and numerically stable particle-in-cell (PIC) simulations of electrostatic plasmas. The starting point is the single-time correlation at the start of the simulation between the statistical fluctuations of the weighted densities of macroparticle centers in the plasma particle phase-space. The fluctuations are associated with different initial conditions, typically due to the random initial conditions (in velocity space) of the macroparticles/simulation plasma, assigned according to their initial distribution of probability. The single-time correlations at all time steps and in each spatial grid cell are then determined from the Laplace–Fourier transforms of themore »« less
https://doi.org/10.1088/1361-6587/ac9016
Uncertainty in Dispersion Forecasting Using Meteorological Ensembles

Conference Leach, M J ; Chin, H-N

A approach for quantifying meteorological uncertainty is via development of an ensemble of forecasts from slightly perturbed initial conditions (Sivillo et al., 1997) to predict the time evolution of the probability density function of atmospheric variables (Mullen and Baurnhefner, 1994). We create an ensemble of forecasts by varying the initial (and boundary) conditions for the COAMPS meteorological model. The variations in the initial conditions must be consistent with analysis error. Optimally, the range of initial conditions would encompass the ''true'' atmospheric state, but which is never actually known. Our method for creating varying initial conditions is to use different globalmore »« less
https://doi.org/10.1007/0-306-47460-3_66

Full Text Available
Time averaging of instantaneous quantities in HYDRA

Technical Report McCallen, R C

For turbulent flow the evaluation of direct numerical simulations (DNS) where all scales are resolved and large-eddy simulation (LES) where only large-scales are resolved is difficult because the results are three-dimensional and transient. To simplify the analysis, the instantaneous flow field can be averaged in time for evaluation and comparison to experimental results. The incompressible Navier-Stokes flow code HYDRA has been modified for calculation of time-average quantities for both DNS and LES. This report describes how time averages of instantaneous quantities are generated during program execution (i.e., while generating the instantaneous quantities, instead of as a postprocessing operation). The calculationsmore »« less
https://doi.org/10.2172/453332

Full Text Available

Similar Records