Analysis of the tradeoffs between energy and run time for multilevel checkpointing

Prasanna Balaprakash; Leonardo A.Bautista Gomez; Mohamed Slim Bouguerra; Stefan M. Wild; Franck Cappello; Paul D. Hovland

doi:10.1007/978-3-319-17248-4_13

Analysis of the tradeoffs between energy and run time for multilevel checkpointing

Prasanna Balaprakash, Leonardo A.Bautista Gomez, Mohamed Slim Bouguerra, Stefan M. Wild, Franck Cappello, Paul D. Hovland

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

5 Scopus citations

Abstract

In high-performance computing, there is a perpetual hunt for performance and scalability. Supercomputers grow larger offering improved computational science throughput. Nevertheless, with an increase in the number of systems’ components and their interactions, the number of failures and the power consumption will increase rapidly. Energy and reliability are among the most challenging issues that need to be addressed for extreme scale computing. We develop analytical models for run time and energy usage for multilevel fault-tolerance schemes. We use these models to study the tradeoff between run time and energy in FTI, a recently developed multilevel checkpoint library, on an IBM Blue Gene/Q. Our results show that energy consumed by FTI is low and the tradeoff between the run time and energy is small. Using the analytical models, we explore the impact of various system-level parameters on run time and energy tradeoffs.

Original language	English
Title of host publication	High Performance Computing Systems
Subtitle of host publication	Performance Modeling, Benchmarking, and Simulation - 5th International Workshop, PMBS 2014, Revised Selected Papers
Editors	Simon D. Hammond, Stephen A. Jarvis, Steven A. Wright
Publisher	Springer Verlag
Pages	249-263
Number of pages	15
ISBN (Print)	9783319172477
DOIs	https://doi.org/10.1007/978-3-319-17248-4_13
State	Published - 2015
Externally published	Yes
Event	5th International Workshop on Performance Modeling, Benchmarking, and Simulation of High Performance Computing Systems, PMBS 2014 - New Orleans, United States Duration: Nov 16 2014 → Nov 16 2014

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	8966
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	5th International Workshop on Performance Modeling, Benchmarking, and Simulation of High Performance Computing Systems, PMBS 2014
Country/Territory	United States
City	New Orleans
Period	11/16/14 → 11/16/14

Funding

This work was supported by the SciDAC and X-Stack activities within the U.S. Department of Energy, Office of Science, Advanced Scientific Computing Research program under contract number DE-AC02-06CH11357.

Access to Document

10.1007/978-3-319-17248-4_13

Cite this

Balaprakash, P., Gomez, L. A. B., Bouguerra, M. S., Wild, S. M., Cappello, F., & Hovland, P. D. (2015). Analysis of the tradeoffs between energy and run time for multilevel checkpointing. In S. D. Hammond, S. A. Jarvis, & S. A. Wright (Eds.), High Performance Computing Systems: Performance Modeling, Benchmarking, and Simulation - 5th International Workshop, PMBS 2014, Revised Selected Papers (pp. 249-263). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8966). Springer Verlag. https://doi.org/10.1007/978-3-319-17248-4_13

Balaprakash, Prasanna ; Gomez, Leonardo A.Bautista ; Bouguerra, Mohamed Slim et al. / Analysis of the tradeoffs between energy and run time for multilevel checkpointing. High Performance Computing Systems: Performance Modeling, Benchmarking, and Simulation - 5th International Workshop, PMBS 2014, Revised Selected Papers. editor / Simon D. Hammond ; Stephen A. Jarvis ; Steven A. Wright. Springer Verlag, 2015. pp. 249-263 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{3d9a4568559a40ed8d21966c9a4ef255,

title = "Analysis of the tradeoffs between energy and run time for multilevel checkpointing",

abstract = "In high-performance computing, there is a perpetual hunt for performance and scalability. Supercomputers grow larger offering improved computational science throughput. Nevertheless, with an increase in the number of systems{\textquoteright} components and their interactions, the number of failures and the power consumption will increase rapidly. Energy and reliability are among the most challenging issues that need to be addressed for extreme scale computing. We develop analytical models for run time and energy usage for multilevel fault-tolerance schemes. We use these models to study the tradeoff between run time and energy in FTI, a recently developed multilevel checkpoint library, on an IBM Blue Gene/Q. Our results show that energy consumed by FTI is low and the tradeoff between the run time and energy is small. Using the analytical models, we explore the impact of various system-level parameters on run time and energy tradeoffs.",

author = "Prasanna Balaprakash and Gomez, {Leonardo A.Bautista} and Bouguerra, {Mohamed Slim} and Wild, {Stefan M.} and Franck Cappello and Hovland, {Paul D.}",

note = "Publisher Copyright: {\textcopyright} Springer International Publishing Switzerland 2015.; 5th International Workshop on Performance Modeling, Benchmarking, and Simulation of High Performance Computing Systems, PMBS 2014 ; Conference date: 16-11-2014 Through 16-11-2014",

year = "2015",

doi = "10.1007/978-3-319-17248-4_13",

language = "English",

isbn = "9783319172477",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "249--263",

editor = "Hammond, {Simon D.} and Jarvis, {Stephen A.} and Wright, {Steven A.}",

booktitle = "High Performance Computing Systems",

}

Balaprakash, P, Gomez, LAB, Bouguerra, MS, Wild, SM, Cappello, F & Hovland, PD 2015, Analysis of the tradeoffs between energy and run time for multilevel checkpointing. in SD Hammond, SA Jarvis & SA Wright (eds), High Performance Computing Systems: Performance Modeling, Benchmarking, and Simulation - 5th International Workshop, PMBS 2014, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 8966, Springer Verlag, pp. 249-263, 5th International Workshop on Performance Modeling, Benchmarking, and Simulation of High Performance Computing Systems, PMBS 2014, New Orleans, United States, 11/16/14. https://doi.org/10.1007/978-3-319-17248-4_13

Analysis of the tradeoffs between energy and run time for multilevel checkpointing. / Balaprakash, Prasanna; Gomez, Leonardo A.Bautista; Bouguerra, Mohamed Slim et al.
High Performance Computing Systems: Performance Modeling, Benchmarking, and Simulation - 5th International Workshop, PMBS 2014, Revised Selected Papers. ed. / Simon D. Hammond; Stephen A. Jarvis; Steven A. Wright. Springer Verlag, 2015. p. 249-263 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8966).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Analysis of the tradeoffs between energy and run time for multilevel checkpointing

AU - Balaprakash, Prasanna

AU - Gomez, Leonardo A.Bautista

AU - Bouguerra, Mohamed Slim

AU - Wild, Stefan M.

AU - Cappello, Franck

AU - Hovland, Paul D.

N1 - Publisher Copyright: © Springer International Publishing Switzerland 2015.

PY - 2015

Y1 - 2015

N2 - In high-performance computing, there is a perpetual hunt for performance and scalability. Supercomputers grow larger offering improved computational science throughput. Nevertheless, with an increase in the number of systems’ components and their interactions, the number of failures and the power consumption will increase rapidly. Energy and reliability are among the most challenging issues that need to be addressed for extreme scale computing. We develop analytical models for run time and energy usage for multilevel fault-tolerance schemes. We use these models to study the tradeoff between run time and energy in FTI, a recently developed multilevel checkpoint library, on an IBM Blue Gene/Q. Our results show that energy consumed by FTI is low and the tradeoff between the run time and energy is small. Using the analytical models, we explore the impact of various system-level parameters on run time and energy tradeoffs.

AB - In high-performance computing, there is a perpetual hunt for performance and scalability. Supercomputers grow larger offering improved computational science throughput. Nevertheless, with an increase in the number of systems’ components and their interactions, the number of failures and the power consumption will increase rapidly. Energy and reliability are among the most challenging issues that need to be addressed for extreme scale computing. We develop analytical models for run time and energy usage for multilevel fault-tolerance schemes. We use these models to study the tradeoff between run time and energy in FTI, a recently developed multilevel checkpoint library, on an IBM Blue Gene/Q. Our results show that energy consumed by FTI is low and the tradeoff between the run time and energy is small. Using the analytical models, we explore the impact of various system-level parameters on run time and energy tradeoffs.

UR - http://www.scopus.com/inward/record.url?scp=84942521141&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-17248-4_13

DO - 10.1007/978-3-319-17248-4_13

M3 - Conference contribution

AN - SCOPUS:84942521141

SN - 9783319172477

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 249

EP - 263

BT - High Performance Computing Systems

A2 - Hammond, Simon D.

A2 - Jarvis, Stephen A.

A2 - Wright, Steven A.

PB - Springer Verlag

T2 - 5th International Workshop on Performance Modeling, Benchmarking, and Simulation of High Performance Computing Systems, PMBS 2014

Y2 - 16 November 2014 through 16 November 2014

ER -

Balaprakash P, Gomez LAB, Bouguerra MS, Wild SM, Cappello F, Hovland PD. Analysis of the tradeoffs between energy and run time for multilevel checkpointing. In Hammond SD, Jarvis SA, Wright SA, editors, High Performance Computing Systems: Performance Modeling, Benchmarking, and Simulation - 5th International Workshop, PMBS 2014, Revised Selected Papers. Springer Verlag. 2015. p. 249-263. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-17248-4_13

Analysis of the tradeoffs between energy and run time for multilevel checkpointing

Abstract

Publication series

Conference

Funding

Access to Document

Other files and links

Fingerprint

Cite this