Performance Evaluation and Energy Consumption for DVFS Processor

Youssef Ait El Mahjoub; Jean Michel Fourneau; Hind Castel-Taleb

doi:10.1007/978-3-030-91825-5_10

Performance Evaluation and Energy Consumption for DVFS Processor

Youssef Ait El Mahjoub
, Jean Michel Fourneau
, Hind Castel-Taleb

Université Versailles-Saint Quentin
Institut Polytechnique de Paris

Résultats de recherche: Le chapitre dans un livre, un rapport, une anthologie ou une collection › Contribution à une conférence › Revue par des pairs

Résumé

Dynamic Voltage and Frequency Scaling (DVFS) involves adjusting the speed to match power consumption and performance requirements of the system. This technology has been shown to provide major power and energy savings in many system components (processor cores, memory system..). In this work, we study the DVFS mechanism for a processor that implements various levels of speed and power (we call these levels “Pstates”). We model the system as a birth-death process and we compare different configurations, using stochastic comparison, in order to evaluate the impact on the response time and power (and energy) consumption. In the case of two Pstates, we proposed a closed form for the steady state distribution and we derived a cost function that uses both performance and energy consumption. Finally, we derive an algorithm that suggests a threshold which minimizes the cost function.

langue originale	Anglais
titre	Performance Engineering and Stochastic Modeling - 17th European Workshop, EPEW 2021, and 26th International Conference, ASMTA 2021, Proceedings
rédacteurs en chef	Paolo Ballarini, Hind Castel, Ioannis Dimitriou, Mauro Iacono, Tuan Phung-Duc, Joris Walraevens
Editeur	Springer Science and Business Media Deutschland GmbH
Pages	165-180
Nombre de pages	16
ISBN (imprimé)	9783030918248
Les DOIs	https://doi.org/10.1007/978-3-030-91825-5_10
état	Publié - 1 janv. 2021
Modification externe	Oui
Evénement	17th European Performance Engineering Workshop, EPEW 2021, and the 26th International Conference on Analytical and Stochastic Modelling Techniques and Applications, ASMTA 2021 - Virtual, Online Durée: 13 déc. 2021 → 14 déc. 2021

Série de publications

Nom	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	13104 LNCS
ISSN (imprimé)	0302-9743
ISSN (Electronique)	1611-3349

Une conférence

Une conférence	17th European Performance Engineering Workshop, EPEW 2021, and the 26th International Conference on Analytical and Stochastic Modelling Techniques and Applications, ASMTA 2021
La ville	Virtual, Online
période	13/12/21 → 14/12/21

SDG des Nations Unies

Ce résultat contribue à ou aux Objectifs de développement durable suivants

SDG 7 Énergie abordable et propre

Accès au document

10.1007/978-3-030-91825-5_10

Contient cette citation

Ait El Mahjoub, Y., Fourneau, J. M., & Castel-Taleb, H. (2021). Performance Evaluation and Energy Consumption for DVFS Processor. Dans P. Ballarini, H. Castel, I. Dimitriou, M. Iacono, T. Phung-Duc, & J. Walraevens (eds.), Performance Engineering and Stochastic Modeling - 17th European Workshop, EPEW 2021, and 26th International Conference, ASMTA 2021, Proceedings (p. 165-180). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol 13104 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-91825-5_10

Ait El Mahjoub, Youssef ; Fourneau, Jean Michel ; Castel-Taleb, Hind. / Performance Evaluation and Energy Consumption for DVFS Processor. Performance Engineering and Stochastic Modeling - 17th European Workshop, EPEW 2021, and 26th International Conference, ASMTA 2021, Proceedings. Editeur / Paolo Ballarini ; Hind Castel ; Ioannis Dimitriou ; Mauro Iacono ; Tuan Phung-Duc ; Joris Walraevens. Springer Science and Business Media Deutschland GmbH, 2021. p. 165-180 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{db61d6dfabac45c897052af7e2cd9a40,

title = "Performance Evaluation and Energy Consumption for DVFS Processor",

abstract = "Dynamic Voltage and Frequency Scaling (DVFS) involves adjusting the speed to match power consumption and performance requirements of the system. This technology has been shown to provide major power and energy savings in many system components (processor cores, memory system..). In this work, we study the DVFS mechanism for a processor that implements various levels of speed and power (we call these levels “Pstates”). We model the system as a birth-death process and we compare different configurations, using stochastic comparison, in order to evaluate the impact on the response time and power (and energy) consumption. In the case of two Pstates, we proposed a closed form for the steady state distribution and we derived a cost function that uses both performance and energy consumption. Finally, we derive an algorithm that suggests a threshold which minimizes the cost function.",

keywords = "DVFS, Energy consumption, Performance evaluation, Stochastic comparison",

author = "\{Ait El Mahjoub\}, Youssef and Fourneau, \{Jean Michel\} and Hind Castel-Taleb",

note = "Publisher Copyright: {\textcopyright} 2021, Springer Nature Switzerland AG.; 17th European Performance Engineering Workshop, EPEW 2021, and the 26th International Conference on Analytical and Stochastic Modelling Techniques and Applications, ASMTA 2021 ; Conference date: 13-12-2021 Through 14-12-2021",

year = "2021",

month = jan,

day = "1",

doi = "10.1007/978-3-030-91825-5\_10",

language = "English",

isbn = "9783030918248",

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

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "165--180",

editor = "Paolo Ballarini and Hind Castel and Ioannis Dimitriou and Mauro Iacono and Tuan Phung-Duc and Joris Walraevens",

booktitle = "Performance Engineering and Stochastic Modeling - 17th European Workshop, EPEW 2021, and 26th International Conference, ASMTA 2021, Proceedings",

}

Ait El Mahjoub, Y, Fourneau, JM & Castel-Taleb, H 2021, Performance Evaluation and Energy Consumption for DVFS Processor. Dans P Ballarini, H Castel, I Dimitriou, M Iacono, T Phung-Duc & J Walraevens (eds), Performance Engineering and Stochastic Modeling - 17th European Workshop, EPEW 2021, and 26th International Conference, ASMTA 2021, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), VOL. 13104 LNCS, Springer Science and Business Media Deutschland GmbH, p. 165-180, 17th European Performance Engineering Workshop, EPEW 2021, and the 26th International Conference on Analytical and Stochastic Modelling Techniques and Applications, ASMTA 2021, Virtual, Online, 13/12/21. https://doi.org/10.1007/978-3-030-91825-5_10

Performance Evaluation and Energy Consumption for DVFS Processor. / Ait El Mahjoub, Youssef; Fourneau, Jean Michel; Castel-Taleb, Hind.
Performance Engineering and Stochastic Modeling - 17th European Workshop, EPEW 2021, and 26th International Conference, ASMTA 2021, Proceedings. Ed. / Paolo Ballarini; Hind Castel; Ioannis Dimitriou; Mauro Iacono; Tuan Phung-Duc; Joris Walraevens. Springer Science and Business Media Deutschland GmbH, 2021. p. 165-180 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol 13104 LNCS).

Résultats de recherche: Le chapitre dans un livre, un rapport, une anthologie ou une collection › Contribution à une conférence › Revue par des pairs

TY - GEN

T1 - Performance Evaluation and Energy Consumption for DVFS Processor

AU - Ait El Mahjoub, Youssef

AU - Fourneau, Jean Michel

AU - Castel-Taleb, Hind

PY - 2021/1/1

Y1 - 2021/1/1

N2 - Dynamic Voltage and Frequency Scaling (DVFS) involves adjusting the speed to match power consumption and performance requirements of the system. This technology has been shown to provide major power and energy savings in many system components (processor cores, memory system..). In this work, we study the DVFS mechanism for a processor that implements various levels of speed and power (we call these levels “Pstates”). We model the system as a birth-death process and we compare different configurations, using stochastic comparison, in order to evaluate the impact on the response time and power (and energy) consumption. In the case of two Pstates, we proposed a closed form for the steady state distribution and we derived a cost function that uses both performance and energy consumption. Finally, we derive an algorithm that suggests a threshold which minimizes the cost function.

AB - Dynamic Voltage and Frequency Scaling (DVFS) involves adjusting the speed to match power consumption and performance requirements of the system. This technology has been shown to provide major power and energy savings in many system components (processor cores, memory system..). In this work, we study the DVFS mechanism for a processor that implements various levels of speed and power (we call these levels “Pstates”). We model the system as a birth-death process and we compare different configurations, using stochastic comparison, in order to evaluate the impact on the response time and power (and energy) consumption. In the case of two Pstates, we proposed a closed form for the steady state distribution and we derived a cost function that uses both performance and energy consumption. Finally, we derive an algorithm that suggests a threshold which minimizes the cost function.

KW - DVFS

KW - Energy consumption

KW - Performance evaluation

KW - Stochastic comparison

U2 - 10.1007/978-3-030-91825-5_10

DO - 10.1007/978-3-030-91825-5_10

M3 - Conference contribution

AN - SCOPUS:85121924258

SN - 9783030918248

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

SP - 165

EP - 180

BT - Performance Engineering and Stochastic Modeling - 17th European Workshop, EPEW 2021, and 26th International Conference, ASMTA 2021, Proceedings

A2 - Ballarini, Paolo

A2 - Castel, Hind

A2 - Dimitriou, Ioannis

A2 - Iacono, Mauro

A2 - Phung-Duc, Tuan

A2 - Walraevens, Joris

PB - Springer Science and Business Media Deutschland GmbH

T2 - 17th European Performance Engineering Workshop, EPEW 2021, and the 26th International Conference on Analytical and Stochastic Modelling Techniques and Applications, ASMTA 2021

Y2 - 13 December 2021 through 14 December 2021

ER -

Ait El Mahjoub Y, Fourneau JM, Castel-Taleb H. Performance Evaluation and Energy Consumption for DVFS Processor. Dans Ballarini P, Castel H, Dimitriou I, Iacono M, Phung-Duc T, Walraevens J, rédacteurs en chef, Performance Engineering and Stochastic Modeling - 17th European Workshop, EPEW 2021, and 26th International Conference, ASMTA 2021, Proceedings. Springer Science and Business Media Deutschland GmbH. 2021. p. 165-180. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-91825-5_10

Performance Evaluation and Energy Consumption for DVFS Processor

Résumé

Série de publications

Une conférence

SDG des Nations Unies

Accès au document

Empreinte digitale

Contient cette citation