Exploring dynamic load imbalance solutions with the CoMD proxy application
Abstract
Proxy applications are developed to simplify studying parallel performance of scientific simulations and to test potential solutions for performance problems. However, proxy applications are typically too simple to allow work migration or to represent the load imbalance of their parent applications. To study the ability of load balancing solutions to balance work effectively, we enable work migration in one of the Exascale Co-design Center for Materials in Extreme Environments (ExMatEx) [1] applications, CoMD. We design a methodology to parameterize three key aspects necessary for studying load imbalance correction: (1) the granularity with which work can be migrated; (2) the initial load imbalance; (3) the dynamic load imbalance (how quickly the load changes over time). Finally, we present a study of the impact of flexibility in work migration in CoMD on load balance and the associated rebalancing costs for a wide range of initial and dynamic load imbalance scenarios.
- Authors:
-
- Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
- Univ. of Alabama, Birmingham, AL (United States). Dept. of Computer and Information Sciences
- Univ. of Copenhagen (Denmark). Dept. of Computer Science
- Publication Date:
- Research Org.:
- Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)
- Sponsoring Org.:
- USDOE National Nuclear Security Administration (NNSA)
- OSTI Identifier:
- 1497267
- Alternate Identifier(s):
- OSTI ID: 1702551
- Report Number(s):
- LLNL-JRNL-725317
Journal ID: ISSN 0167-739X; 874899
- Grant/Contract Number:
- AC52-07NA27344; LLNL-JRNL-725317
- Resource Type:
- Accepted Manuscript
- Journal Name:
- Future Generations Computer Systems
- Additional Journal Information:
- Journal Volume: 92; Journal Issue: C; Journal ID: ISSN 0167-739X
- Publisher:
- Elsevier
- Country of Publication:
- United States
- Language:
- English
- Subject:
- 97 MATHEMATICS AND COMPUTING
Citation Formats
Pearce, Olga, Ahmed, Hadia, Larsen, Rasmus W., Pirkelbauer, Peter, and Richards, David F. Exploring dynamic load imbalance solutions with the CoMD proxy application. United States: N. p., 2017.
Web. doi:10.1016/j.future.2017.12.010.
Pearce, Olga, Ahmed, Hadia, Larsen, Rasmus W., Pirkelbauer, Peter, & Richards, David F. Exploring dynamic load imbalance solutions with the CoMD proxy application. United States. https://doi.org/10.1016/j.future.2017.12.010
Pearce, Olga, Ahmed, Hadia, Larsen, Rasmus W., Pirkelbauer, Peter, and Richards, David F. Wed .
"Exploring dynamic load imbalance solutions with the CoMD proxy application". United States. https://doi.org/10.1016/j.future.2017.12.010. https://www.osti.gov/servlets/purl/1497267.
@article{osti_1497267,
title = {Exploring dynamic load imbalance solutions with the CoMD proxy application},
author = {Pearce, Olga and Ahmed, Hadia and Larsen, Rasmus W. and Pirkelbauer, Peter and Richards, David F.},
abstractNote = {Proxy applications are developed to simplify studying parallel performance of scientific simulations and to test potential solutions for performance problems. However, proxy applications are typically too simple to allow work migration or to represent the load imbalance of their parent applications. To study the ability of load balancing solutions to balance work effectively, we enable work migration in one of the Exascale Co-design Center for Materials in Extreme Environments (ExMatEx) [1] applications, CoMD. We design a methodology to parameterize three key aspects necessary for studying load imbalance correction: (1) the granularity with which work can be migrated; (2) the initial load imbalance; (3) the dynamic load imbalance (how quickly the load changes over time). Finally, we present a study of the impact of flexibility in work migration in CoMD on load balance and the associated rebalancing costs for a wide range of initial and dynamic load imbalance scenarios.},
doi = {10.1016/j.future.2017.12.010},
journal = {Future Generations Computer Systems},
number = C,
volume = 92,
place = {United States},
year = {Wed Dec 20 00:00:00 EST 2017},
month = {Wed Dec 20 00:00:00 EST 2017}
}
Web of Science
Figures / Tables:
Figures / Tables found in this record: