Dynamic Power Management using adaptive learning tree

Eui Young Chung, Luca Benini, Giovanni De Micheli

Research output: Contribution to journalConference article

128 Citations (Scopus)

Abstract

Dynamic Power Management (DPM) is a technique to reduce power consumption of electronic systems by selectively shutting down idle components. The quality of the shutdown control algorithm (power management policy) mostly depends on the knowledge of user behavior, which in many cases is initially unknown or non-stationary. For this reason, DPM policies should be capable of adapting to changes in user behavior. In this paper, we present a novel DPM scheme based on idle period clustering and adaptive learning trees. We also provide a design guide for applying our technique to components with multiple sleep states. Experimental results show that our technique outperforms other advanced DPM schemes as well as simple time-out policies. The proposed approach shows little deviation of efficiency for various workloads having different characteristics, while other policies show that their efficiency changes drastically depending on the trace data characteristics. Furthermore, experimental evidence indicates that our workload learning algorithm is stable and has fast convergence.

Original languageEnglish
Pages (from-to)274-279
Number of pages6
JournalIEEE/ACM International Conference on Computer-Aided Design, Digest of Technical Papers
Publication statusPublished - 1999 Dec 1
EventProceedings of the 1999 IEEE/ACM International Conference on Computer-Aided Design (ICCAD-99) - San Jose, CA, USA
Duration: 1999 Nov 71999 Nov 11

Fingerprint

Learning algorithms
Electric power utilization
Power management
Sleep

All Science Journal Classification (ASJC) codes

  • Software
  • Computer Science Applications
  • Computer Graphics and Computer-Aided Design

Cite this

@article{822ed3af8d2040d68b8adb8df973e68d,
title = "Dynamic Power Management using adaptive learning tree",
abstract = "Dynamic Power Management (DPM) is a technique to reduce power consumption of electronic systems by selectively shutting down idle components. The quality of the shutdown control algorithm (power management policy) mostly depends on the knowledge of user behavior, which in many cases is initially unknown or non-stationary. For this reason, DPM policies should be capable of adapting to changes in user behavior. In this paper, we present a novel DPM scheme based on idle period clustering and adaptive learning trees. We also provide a design guide for applying our technique to components with multiple sleep states. Experimental results show that our technique outperforms other advanced DPM schemes as well as simple time-out policies. The proposed approach shows little deviation of efficiency for various workloads having different characteristics, while other policies show that their efficiency changes drastically depending on the trace data characteristics. Furthermore, experimental evidence indicates that our workload learning algorithm is stable and has fast convergence.",
author = "Chung, {Eui Young} and Luca Benini and {De Micheli}, Giovanni",
year = "1999",
month = "12",
day = "1",
language = "English",
pages = "274--279",
journal = "IEEE/ACM International Conference on Computer-Aided Design, Digest of Technical Papers",
issn = "1092-3152",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

Dynamic Power Management using adaptive learning tree. / Chung, Eui Young; Benini, Luca; De Micheli, Giovanni.

In: IEEE/ACM International Conference on Computer-Aided Design, Digest of Technical Papers, 01.12.1999, p. 274-279.

Research output: Contribution to journalConference article

TY - JOUR

T1 - Dynamic Power Management using adaptive learning tree

AU - Chung, Eui Young

AU - Benini, Luca

AU - De Micheli, Giovanni

PY - 1999/12/1

Y1 - 1999/12/1

N2 - Dynamic Power Management (DPM) is a technique to reduce power consumption of electronic systems by selectively shutting down idle components. The quality of the shutdown control algorithm (power management policy) mostly depends on the knowledge of user behavior, which in many cases is initially unknown or non-stationary. For this reason, DPM policies should be capable of adapting to changes in user behavior. In this paper, we present a novel DPM scheme based on idle period clustering and adaptive learning trees. We also provide a design guide for applying our technique to components with multiple sleep states. Experimental results show that our technique outperforms other advanced DPM schemes as well as simple time-out policies. The proposed approach shows little deviation of efficiency for various workloads having different characteristics, while other policies show that their efficiency changes drastically depending on the trace data characteristics. Furthermore, experimental evidence indicates that our workload learning algorithm is stable and has fast convergence.

AB - Dynamic Power Management (DPM) is a technique to reduce power consumption of electronic systems by selectively shutting down idle components. The quality of the shutdown control algorithm (power management policy) mostly depends on the knowledge of user behavior, which in many cases is initially unknown or non-stationary. For this reason, DPM policies should be capable of adapting to changes in user behavior. In this paper, we present a novel DPM scheme based on idle period clustering and adaptive learning trees. We also provide a design guide for applying our technique to components with multiple sleep states. Experimental results show that our technique outperforms other advanced DPM schemes as well as simple time-out policies. The proposed approach shows little deviation of efficiency for various workloads having different characteristics, while other policies show that their efficiency changes drastically depending on the trace data characteristics. Furthermore, experimental evidence indicates that our workload learning algorithm is stable and has fast convergence.

UR - http://www.scopus.com/inward/record.url?scp=0033337995&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0033337995&partnerID=8YFLogxK

M3 - Conference article

AN - SCOPUS:0033337995

SP - 274

EP - 279

JO - IEEE/ACM International Conference on Computer-Aided Design, Digest of Technical Papers

JF - IEEE/ACM International Conference on Computer-Aided Design, Digest of Technical Papers

SN - 1092-3152

ER -