An evaluation of code and data optimizations in the context of disk power reduction

Mahmut Kandemir, Seung Woo Son, Guangyu Chen

Research output: Contribution to journalArticle

7 Citations (Scopus)

Abstract

Disk power management is becoming increasingly important in high-end server and cluster type of environments that execute dataintensive applications. While hardware-only approaches (e.g., low-power modes supported by current disks) are successful to a certain extent, one also needs to consider the software side to achieve further energy savings. This paper first demonstrates that conventional data locality oriented code transformations are not sufficient for minimizing disk power consumption. The reason is that these optimizations do not take into account how disk-resident array data are laid out on the disk system, and consequently, fail to increase idle periods of disks, which is the primary metric using which disk power can be reduced. Instead, we propose a disk layout aware application optimization strategy that uses both code restructuring and data layout optimization. Our experimental evaluation with several benchmark codes reveal that the proposed strategy is very successful in reducing disk energy consumption without performing much worse than a pure data locality oriented scheme, as far as execution cycles are concerned. The experiments also show that the benefits coming from our approach increase with the increased number of disks; i.e., it scales very well.

Original languageEnglish (US)
Pages (from-to)209-214
Number of pages6
JournalProceedings of the International Symposium on Low Power Electronics and Design
StatePublished - 2005

Fingerprint

Energy conservation
Electric power utilization
Servers
Energy utilization
Hardware
Experiments
Power management

All Science Journal Classification (ASJC) codes

  • Engineering(all)

Cite this

@article{299b1a8f56844447a21e291f11a469d7,
title = "An evaluation of code and data optimizations in the context of disk power reduction",
abstract = "Disk power management is becoming increasingly important in high-end server and cluster type of environments that execute dataintensive applications. While hardware-only approaches (e.g., low-power modes supported by current disks) are successful to a certain extent, one also needs to consider the software side to achieve further energy savings. This paper first demonstrates that conventional data locality oriented code transformations are not sufficient for minimizing disk power consumption. The reason is that these optimizations do not take into account how disk-resident array data are laid out on the disk system, and consequently, fail to increase idle periods of disks, which is the primary metric using which disk power can be reduced. Instead, we propose a disk layout aware application optimization strategy that uses both code restructuring and data layout optimization. Our experimental evaluation with several benchmark codes reveal that the proposed strategy is very successful in reducing disk energy consumption without performing much worse than a pure data locality oriented scheme, as far as execution cycles are concerned. The experiments also show that the benefits coming from our approach increase with the increased number of disks; i.e., it scales very well.",
author = "Mahmut Kandemir and Son, {Seung Woo} and Guangyu Chen",
year = "2005",
language = "English (US)",
pages = "209--214",
journal = "Proceedings of the International Symposium on Low Power Design",
issn = "1533-4678",

}

TY - JOUR

T1 - An evaluation of code and data optimizations in the context of disk power reduction

AU - Kandemir, Mahmut

AU - Son, Seung Woo

AU - Chen, Guangyu

PY - 2005

Y1 - 2005

N2 - Disk power management is becoming increasingly important in high-end server and cluster type of environments that execute dataintensive applications. While hardware-only approaches (e.g., low-power modes supported by current disks) are successful to a certain extent, one also needs to consider the software side to achieve further energy savings. This paper first demonstrates that conventional data locality oriented code transformations are not sufficient for minimizing disk power consumption. The reason is that these optimizations do not take into account how disk-resident array data are laid out on the disk system, and consequently, fail to increase idle periods of disks, which is the primary metric using which disk power can be reduced. Instead, we propose a disk layout aware application optimization strategy that uses both code restructuring and data layout optimization. Our experimental evaluation with several benchmark codes reveal that the proposed strategy is very successful in reducing disk energy consumption without performing much worse than a pure data locality oriented scheme, as far as execution cycles are concerned. The experiments also show that the benefits coming from our approach increase with the increased number of disks; i.e., it scales very well.

AB - Disk power management is becoming increasingly important in high-end server and cluster type of environments that execute dataintensive applications. While hardware-only approaches (e.g., low-power modes supported by current disks) are successful to a certain extent, one also needs to consider the software side to achieve further energy savings. This paper first demonstrates that conventional data locality oriented code transformations are not sufficient for minimizing disk power consumption. The reason is that these optimizations do not take into account how disk-resident array data are laid out on the disk system, and consequently, fail to increase idle periods of disks, which is the primary metric using which disk power can be reduced. Instead, we propose a disk layout aware application optimization strategy that uses both code restructuring and data layout optimization. Our experimental evaluation with several benchmark codes reveal that the proposed strategy is very successful in reducing disk energy consumption without performing much worse than a pure data locality oriented scheme, as far as execution cycles are concerned. The experiments also show that the benefits coming from our approach increase with the increased number of disks; i.e., it scales very well.

UR - http://www.scopus.com/inward/record.url?scp=28444474890&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=28444474890&partnerID=8YFLogxK

M3 - Article

SP - 209

EP - 214

JO - Proceedings of the International Symposium on Low Power Design

JF - Proceedings of the International Symposium on Low Power Design

SN - 1533-4678

ER -