Just in time dynamic voltage scaling

Exploiting inter-node slack to save energy in MPI programs

Nandini Kappiah, Vincent W. Freeh, David K Lowenthal

Research output: Chapter in Book/Report/Conference proceedingConference contribution

133 Citations (Scopus)

Abstract

Recently, improving the energy efficiency of HPC machines has become important. As a result, interest in using power-scalable clusters, where frequency and voltage can be dynamically modified, has increased. On power-scalable clusters, one opportunity for saving energy with little or no loss of performance exists when the computational load is not perfectly balanced. This situation occurs frequently, as balancing load between nodes is one of the long standing problems in parallel and distributed computing. In this paper we present a system called Jitter, which reduces the frequency on nodes that are assigned less computation and therefore have slack time. This saves energy on these nodes, and the goal of Jitter is to attempt to ensure that they arrive "just in time" so that they avoid increasing overall execution time. For example, in Aztec, from the ASCI Purple suite, our algorithm uses 8% less energy while increasing execution time by only 2.6%.

Original languageEnglish (US)
Title of host publicationProceedings of the ACM/IEEE 2005 Supercomputing Conference, SC'05
Volume2005
DOIs
StatePublished - 2005
Externally publishedYes
EventACM/IEEE 2005 Supercomputing Conference, SC'05 - Seatle, WA, United States
Duration: Nov 12 2005Nov 18 2005

Other

OtherACM/IEEE 2005 Supercomputing Conference, SC'05
CountryUnited States
CitySeatle, WA
Period11/12/0511/18/05

Fingerprint

Jitter
Distributed computer systems
Parallel processing systems
Resource allocation
Energy efficiency
Energy conservation
Electric potential
Voltage scaling

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Kappiah, N., Freeh, V. W., & Lowenthal, D. K. (2005). Just in time dynamic voltage scaling: Exploiting inter-node slack to save energy in MPI programs. In Proceedings of the ACM/IEEE 2005 Supercomputing Conference, SC'05 (Vol. 2005). [1559985] https://doi.org/10.1109/SC.2005.39

Just in time dynamic voltage scaling : Exploiting inter-node slack to save energy in MPI programs. / Kappiah, Nandini; Freeh, Vincent W.; Lowenthal, David K.

Proceedings of the ACM/IEEE 2005 Supercomputing Conference, SC'05. Vol. 2005 2005. 1559985.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Kappiah, N, Freeh, VW & Lowenthal, DK 2005, Just in time dynamic voltage scaling: Exploiting inter-node slack to save energy in MPI programs. in Proceedings of the ACM/IEEE 2005 Supercomputing Conference, SC'05. vol. 2005, 1559985, ACM/IEEE 2005 Supercomputing Conference, SC'05, Seatle, WA, United States, 11/12/05. https://doi.org/10.1109/SC.2005.39
Kappiah N, Freeh VW, Lowenthal DK. Just in time dynamic voltage scaling: Exploiting inter-node slack to save energy in MPI programs. In Proceedings of the ACM/IEEE 2005 Supercomputing Conference, SC'05. Vol. 2005. 2005. 1559985 https://doi.org/10.1109/SC.2005.39
Kappiah, Nandini ; Freeh, Vincent W. ; Lowenthal, David K. / Just in time dynamic voltage scaling : Exploiting inter-node slack to save energy in MPI programs. Proceedings of the ACM/IEEE 2005 Supercomputing Conference, SC'05. Vol. 2005 2005.
@inproceedings{d7982a2ceada4d068c98a050be88bb95,
title = "Just in time dynamic voltage scaling: Exploiting inter-node slack to save energy in MPI programs",
abstract = "Recently, improving the energy efficiency of HPC machines has become important. As a result, interest in using power-scalable clusters, where frequency and voltage can be dynamically modified, has increased. On power-scalable clusters, one opportunity for saving energy with little or no loss of performance exists when the computational load is not perfectly balanced. This situation occurs frequently, as balancing load between nodes is one of the long standing problems in parallel and distributed computing. In this paper we present a system called Jitter, which reduces the frequency on nodes that are assigned less computation and therefore have slack time. This saves energy on these nodes, and the goal of Jitter is to attempt to ensure that they arrive {"}just in time{"} so that they avoid increasing overall execution time. For example, in Aztec, from the ASCI Purple suite, our algorithm uses 8{\%} less energy while increasing execution time by only 2.6{\%}.",
author = "Nandini Kappiah and Freeh, {Vincent W.} and Lowenthal, {David K}",
year = "2005",
doi = "10.1109/SC.2005.39",
language = "English (US)",
isbn = "1595930612",
volume = "2005",
booktitle = "Proceedings of the ACM/IEEE 2005 Supercomputing Conference, SC'05",

}

TY - GEN

T1 - Just in time dynamic voltage scaling

T2 - Exploiting inter-node slack to save energy in MPI programs

AU - Kappiah, Nandini

AU - Freeh, Vincent W.

AU - Lowenthal, David K

PY - 2005

Y1 - 2005

N2 - Recently, improving the energy efficiency of HPC machines has become important. As a result, interest in using power-scalable clusters, where frequency and voltage can be dynamically modified, has increased. On power-scalable clusters, one opportunity for saving energy with little or no loss of performance exists when the computational load is not perfectly balanced. This situation occurs frequently, as balancing load between nodes is one of the long standing problems in parallel and distributed computing. In this paper we present a system called Jitter, which reduces the frequency on nodes that are assigned less computation and therefore have slack time. This saves energy on these nodes, and the goal of Jitter is to attempt to ensure that they arrive "just in time" so that they avoid increasing overall execution time. For example, in Aztec, from the ASCI Purple suite, our algorithm uses 8% less energy while increasing execution time by only 2.6%.

AB - Recently, improving the energy efficiency of HPC machines has become important. As a result, interest in using power-scalable clusters, where frequency and voltage can be dynamically modified, has increased. On power-scalable clusters, one opportunity for saving energy with little or no loss of performance exists when the computational load is not perfectly balanced. This situation occurs frequently, as balancing load between nodes is one of the long standing problems in parallel and distributed computing. In this paper we present a system called Jitter, which reduces the frequency on nodes that are assigned less computation and therefore have slack time. This saves energy on these nodes, and the goal of Jitter is to attempt to ensure that they arrive "just in time" so that they avoid increasing overall execution time. For example, in Aztec, from the ASCI Purple suite, our algorithm uses 8% less energy while increasing execution time by only 2.6%.

UR - http://www.scopus.com/inward/record.url?scp=33845438524&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33845438524&partnerID=8YFLogxK

U2 - 10.1109/SC.2005.39

DO - 10.1109/SC.2005.39

M3 - Conference contribution

SN - 1595930612

SN - 9781595930613

VL - 2005

BT - Proceedings of the ACM/IEEE 2005 Supercomputing Conference, SC'05

ER -