A prefetching algorithm for multi-speed disks

Seung Woo Son, Mahmut Kandemir

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

Abstract

Power consumption of disk based storage systems is becoming an increasingly pressing issue for both the commercial and the scientific application domains. Prior work proposed several hardware based approaches to reducing disk power consumption by making use of techniques such as spinning down idle disks and rotating them at lower speeds than the maximum speed possible. While such techniques are certainly very important, it is also critical to consider the influence the software can exercise in shaping the power consumption of disk-intensive application programs. Motivated by this observation, the main goal of this work is to study whether an optimizing compiler can be used for increasing the power benefits that could be obtained from multi-speed disks. Specifically, we propose and experimentally evaluate a compiler-directed energy-aware data prefetching scheme for scientific applications that process disk-resident data sets. This scheme automatically determines the prefetch distances for all disk access instructions, the disk speeds to be employed, and the associated disk layouts (striping parameters) in a unified setting. We implemented the proposed approach within an optimizing compiler framework and conducted experiments with several disk-intensive applications. Our experimental evaluation shows that the proposed approach brings significant reductions in disk energy consumption over a state-of-the-art software-based I/O prefetching mechanism that does not take into account energy consumption explicitly. Our results also show that the energy-aware prefetching scheme does not bring any extra performance penalties and the energy reductions achieved are consistent across a wide spectrum of values of the simulation parameters. We also show that our scheme can be extended to multiple application scenarios using a hierarchical disk layout determination.

Original languageEnglish (US)
Title of host publicationTransactions on High-Performance Embedded Architectures and Compilers I
Pages317-340
Number of pages24
DOIs
StatePublished - Dec 1 2007

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume4050 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Fingerprint

Prefetching
Electric power utilization
Energy utilization
Application programs
Optimizing Compilers
Power Consumption
Hardware
Energy Consumption
Layout
Energy
Software
Experiments
Storage System
Experimental Evaluation
Compiler
Exercise
Penalty
Rotating

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Son, S. W., & Kandemir, M. (2007). A prefetching algorithm for multi-speed disks. In Transactions on High-Performance Embedded Architectures and Compilers I (pp. 317-340). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4050 LNCS). https://doi.org/10.1007/978-3-540-71528-3_20
Son, Seung Woo ; Kandemir, Mahmut. / A prefetching algorithm for multi-speed disks. Transactions on High-Performance Embedded Architectures and Compilers I. 2007. pp. 317-340 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{480f375bb90b4e7a8994462a40b27807,
title = "A prefetching algorithm for multi-speed disks",
abstract = "Power consumption of disk based storage systems is becoming an increasingly pressing issue for both the commercial and the scientific application domains. Prior work proposed several hardware based approaches to reducing disk power consumption by making use of techniques such as spinning down idle disks and rotating them at lower speeds than the maximum speed possible. While such techniques are certainly very important, it is also critical to consider the influence the software can exercise in shaping the power consumption of disk-intensive application programs. Motivated by this observation, the main goal of this work is to study whether an optimizing compiler can be used for increasing the power benefits that could be obtained from multi-speed disks. Specifically, we propose and experimentally evaluate a compiler-directed energy-aware data prefetching scheme for scientific applications that process disk-resident data sets. This scheme automatically determines the prefetch distances for all disk access instructions, the disk speeds to be employed, and the associated disk layouts (striping parameters) in a unified setting. We implemented the proposed approach within an optimizing compiler framework and conducted experiments with several disk-intensive applications. Our experimental evaluation shows that the proposed approach brings significant reductions in disk energy consumption over a state-of-the-art software-based I/O prefetching mechanism that does not take into account energy consumption explicitly. Our results also show that the energy-aware prefetching scheme does not bring any extra performance penalties and the energy reductions achieved are consistent across a wide spectrum of values of the simulation parameters. We also show that our scheme can be extended to multiple application scenarios using a hierarchical disk layout determination.",
author = "Son, {Seung Woo} and Mahmut Kandemir",
year = "2007",
month = "12",
day = "1",
doi = "10.1007/978-3-540-71528-3_20",
language = "English (US)",
isbn = "3540715274",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "317--340",
booktitle = "Transactions on High-Performance Embedded Architectures and Compilers I",

}

Son, SW & Kandemir, M 2007, A prefetching algorithm for multi-speed disks. in Transactions on High-Performance Embedded Architectures and Compilers I. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 4050 LNCS, pp. 317-340. https://doi.org/10.1007/978-3-540-71528-3_20

A prefetching algorithm for multi-speed disks. / Son, Seung Woo; Kandemir, Mahmut.

Transactions on High-Performance Embedded Architectures and Compilers I. 2007. p. 317-340 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4050 LNCS).

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

TY - GEN

T1 - A prefetching algorithm for multi-speed disks

AU - Son, Seung Woo

AU - Kandemir, Mahmut

PY - 2007/12/1

Y1 - 2007/12/1

N2 - Power consumption of disk based storage systems is becoming an increasingly pressing issue for both the commercial and the scientific application domains. Prior work proposed several hardware based approaches to reducing disk power consumption by making use of techniques such as spinning down idle disks and rotating them at lower speeds than the maximum speed possible. While such techniques are certainly very important, it is also critical to consider the influence the software can exercise in shaping the power consumption of disk-intensive application programs. Motivated by this observation, the main goal of this work is to study whether an optimizing compiler can be used for increasing the power benefits that could be obtained from multi-speed disks. Specifically, we propose and experimentally evaluate a compiler-directed energy-aware data prefetching scheme for scientific applications that process disk-resident data sets. This scheme automatically determines the prefetch distances for all disk access instructions, the disk speeds to be employed, and the associated disk layouts (striping parameters) in a unified setting. We implemented the proposed approach within an optimizing compiler framework and conducted experiments with several disk-intensive applications. Our experimental evaluation shows that the proposed approach brings significant reductions in disk energy consumption over a state-of-the-art software-based I/O prefetching mechanism that does not take into account energy consumption explicitly. Our results also show that the energy-aware prefetching scheme does not bring any extra performance penalties and the energy reductions achieved are consistent across a wide spectrum of values of the simulation parameters. We also show that our scheme can be extended to multiple application scenarios using a hierarchical disk layout determination.

AB - Power consumption of disk based storage systems is becoming an increasingly pressing issue for both the commercial and the scientific application domains. Prior work proposed several hardware based approaches to reducing disk power consumption by making use of techniques such as spinning down idle disks and rotating them at lower speeds than the maximum speed possible. While such techniques are certainly very important, it is also critical to consider the influence the software can exercise in shaping the power consumption of disk-intensive application programs. Motivated by this observation, the main goal of this work is to study whether an optimizing compiler can be used for increasing the power benefits that could be obtained from multi-speed disks. Specifically, we propose and experimentally evaluate a compiler-directed energy-aware data prefetching scheme for scientific applications that process disk-resident data sets. This scheme automatically determines the prefetch distances for all disk access instructions, the disk speeds to be employed, and the associated disk layouts (striping parameters) in a unified setting. We implemented the proposed approach within an optimizing compiler framework and conducted experiments with several disk-intensive applications. Our experimental evaluation shows that the proposed approach brings significant reductions in disk energy consumption over a state-of-the-art software-based I/O prefetching mechanism that does not take into account energy consumption explicitly. Our results also show that the energy-aware prefetching scheme does not bring any extra performance penalties and the energy reductions achieved are consistent across a wide spectrum of values of the simulation parameters. We also show that our scheme can be extended to multiple application scenarios using a hierarchical disk layout determination.

UR - http://www.scopus.com/inward/record.url?scp=38549173741&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=38549173741&partnerID=8YFLogxK

U2 - 10.1007/978-3-540-71528-3_20

DO - 10.1007/978-3-540-71528-3_20

M3 - Conference contribution

SN - 3540715274

SN - 9783540715276

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

SP - 317

EP - 340

BT - Transactions on High-Performance Embedded Architectures and Compilers I

ER -

Son SW, Kandemir M. A prefetching algorithm for multi-speed disks. In Transactions on High-Performance Embedded Architectures and Compilers I. 2007. p. 317-340. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-540-71528-3_20