High-order asynchrony-tolerant finite difference schemes for partial differential equations
Abstract
Synchronizations of processing elements (PEs) in massively parallel simulations, which arise due to communication or load imbalances between PEs, significantly affect the scalability of scientific applications. We have recently proposed a method based on finite-difference schemes to solve partial differential equations in an asynchronous fashion – synchronization between PEs is relaxed at a mathematical level. While standard schemes can maintain their stability in the presence of asynchrony, their accuracy is drastically affected. In this work, we present a general methodology to derive asynchrony-tolerant (AT) finite difference schemes of arbitrary order of accuracy, which can maintain their accuracy when synchronizations are relaxed. We show that there are several choices available in selecting a stencil to derive these schemes and discuss their effect on numerical and computational performance. We provide a simple classification of schemes based on the stencil and derive schemes that are representative of different classes. Their numerical error is rigorously analyzed within a statistical framework to obtain the overall accuracy of the solution. Results from numerical experiments are used to validate the performance of the schemes.
- Authors:
-
- Texas A&M University, College Station, TX (United States). Department of Aerospace Engineering
- Publication Date:
- Research Org.:
- Lawrence Berkeley National Laboratory (LBNL), Berkeley, CA (United States). National Energy Research Scientific Computing Center (NERSC)
- Sponsoring Org.:
- USDOE Office of Science (SC), Advanced Scientific Computing Research (ASCR)
- OSTI Identifier:
- 1463649
- Resource Type:
- Journal Article: Accepted Manuscript
- Journal Name:
- Journal of Computational Physics
- Additional Journal Information:
- Journal Volume: 350; Journal Issue: C; Journal ID: ISSN 0021-9991
- Publisher:
- Elsevier
- Country of Publication:
- United States
- Language:
- English
- Subject:
- 42 ENGINEERING
Citation Formats
Aditya, Konduri, and Donzis, Diego A. High-order asynchrony-tolerant finite difference schemes for partial differential equations. United States: N. p., 2017.
Web. doi:10.1016/j.jcp.2017.08.037.
Aditya, Konduri, & Donzis, Diego A. High-order asynchrony-tolerant finite difference schemes for partial differential equations. United States. https://doi.org/10.1016/j.jcp.2017.08.037
Aditya, Konduri, and Donzis, Diego A. 2017.
"High-order asynchrony-tolerant finite difference schemes for partial differential equations". United States. https://doi.org/10.1016/j.jcp.2017.08.037. https://www.osti.gov/servlets/purl/1463649.
@article{osti_1463649,
title = {High-order asynchrony-tolerant finite difference schemes for partial differential equations},
author = {Aditya, Konduri and Donzis, Diego A.},
abstractNote = {Synchronizations of processing elements (PEs) in massively parallel simulations, which arise due to communication or load imbalances between PEs, significantly affect the scalability of scientific applications. We have recently proposed a method based on finite-difference schemes to solve partial differential equations in an asynchronous fashion – synchronization between PEs is relaxed at a mathematical level. While standard schemes can maintain their stability in the presence of asynchrony, their accuracy is drastically affected. In this work, we present a general methodology to derive asynchrony-tolerant (AT) finite difference schemes of arbitrary order of accuracy, which can maintain their accuracy when synchronizations are relaxed. We show that there are several choices available in selecting a stencil to derive these schemes and discuss their effect on numerical and computational performance. We provide a simple classification of schemes based on the stencil and derive schemes that are representative of different classes. Their numerical error is rigorously analyzed within a statistical framework to obtain the overall accuracy of the solution. Results from numerical experiments are used to validate the performance of the schemes.},
doi = {10.1016/j.jcp.2017.08.037},
url = {https://www.osti.gov/biblio/1463649},
journal = {Journal of Computational Physics},
issn = {0021-9991},
number = C,
volume = 350,
place = {United States},
year = {Fri Dec 01 00:00:00 EST 2017},
month = {Fri Dec 01 00:00:00 EST 2017}
}
Web of Science
Works referenced in this record:
The Turbulent Schmidt Number
journal, April 2014
- Donzis, Diego A.; Aditya, Konduri; Sreenivasan, K. R.
- Journal of Fluids Engineering, Vol. 136, Issue 6
Poster: Asynchronous Computing for Partial Differential Equations at Extreme Scales
conference, November 2012
- Konduri, Aditya; Donzis, Diego A.
- 2012 SC Companion: High Performance Computing, Networking Storage and Analysis
Asynchronous finite-difference schemes for partial differential equations
journal, October 2014
- Donzis, Diego A.; Aditya, Konduri
- Journal of Computational Physics, Vol. 274
Fluctuations of thermodynamic variables in stationary compressible turbulence
journal, September 2013
- Donzis, Diego A.; Jagannathan, Shriram
- Journal of Fluid Mechanics, Vol. 733
The Turbulent Schmidt Number
journal, April 2014
- Donzis, Diego A.; Aditya, Konduri; Sreenivasan, K. R.
- Journal of Fluids Engineering, Vol. 136, Issue 6
DNS of a turbulent lifted DME jet flame
journal, July 2016
- Minamoto, Yuki; Chen, Jacqueline H.
- Combustion and Flame, Vol. 169
The International Exascale Software Project roadmap
journal, January 2011
- Dongarra, Jack; Beckman, Pete; Moore, Terry
- The International Journal of High Performance Computing Applications, Vol. 25, Issue 1
Communication-Avoiding Symmetric-Indefinite Factorization
journal, January 2014
- Ballard, Grey; Becker, Dulceneia; Demmel, James
- SIAM Journal on Matrix Analysis and Applications, Vol. 35, Issue 4
Asynchronous and corrected-asynchronous finite difference solutions of PDEs on MIMD multiprocessors
journal, September 1994
- Amitai, Dganit; Averbuch, Amir; Itzikowitz, Samuel
- Numerical Algorithms, Vol. 6, Issue 2
On parallel asynchronous high-order solutions of parabolic PDEs
journal, March 1996
- Amitai, Dganit; Averbuch, Amir; Israeli, Moshe
- Numerical Algorithms, Vol. 12, Issue 1
Asynchronous finite-difference schemes for partial differential equations
journal, October 2014
- Donzis, Diego A.; Aditya, Konduri
- Journal of Computational Physics, Vol. 274
Delayed Difference Scheme for Large Scale Scientific Simulations
journal, November 2014
- Mudigere, Dheevatsa; Sherlekar, Sunil D.; Ansumali, Santosh
- Physical Review Letters, Vol. 113, Issue 21
The Phenomenology of Small-Scale Turbulence
journal, January 1997
- Sreenivasan, K. R.; Antonia, R. A.
- Annual Review of Fluid Mechanics, Vol. 29, Issue 1
Review of summation-by-parts schemes for initial–boundary-value problems
journal, July 2014
- Svärd, Magnus; Nordström, Jan
- Journal of Computational Physics, Vol. 268
Works referencing / citing this record:
Direct Numerical Simulations of turbulent flows using high-order Asynchrony-Tolerant schemes: accuracy and performance
text, January 2020
- Kumari, Komal; Donzis, Diego A.
- arXiv
A scalable weakly-synchronous algorithm for solving partial differential equations
preprint, January 2019
- Aditya, Konduri; Gysi, Tobias; Kwasniewski, Grzegorz
- arXiv