Workload clustering for increasing energy savings on embedded MPSoCs

S. H.K. Narayanan, O. Ozturk, Mahmut Kandemir, M. Karakoy

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

4 Citations (Scopus)

Abstract

Voltage/frequency scaling and processor low-power modes (i. e., processor shut-down) are two important mechanisms used for reducing energy consumption in embedded MPSoCs. While a unified scheme that combines these two mechanisms can achieve significant savings in some cases, such an approach is limited by the code parallelization strategy employed. In this paper, we propose a novel, integer linear programming (ILP) based workload clustering strategy across parallel processors, oriented towards maximizing the number of idle processors without impacting original execution times. These idle processors can then be switched to a low power mode to maximize energy savings, whereas the remaining ones can make use of voltage/frequency scaling. In order to check whether this approach brings any energy benefits over the pure voltage scaling based, pure processor shut-down based, or a simple unified scheme, we implemented four different approaches and tested them using a set of eight array/loop-intensive embedded applications. Our simulation-based analysis reveals that the proposed ILP based approach (1) is very effective in reducing the energy consumptions of the applications tested and (2) generates much better energy savings than all the alternate schemes tested (including a unified scheme that combines voltage/frequency scaling and processor shutdown).

Original languageEnglish (US)
Title of host publicationProceedings - IEEE International SOC Conference, 2005 SOCC
EditorsD. Ha, R. Krishnamurthy, S. Kim, A. Marshall
Pages157-160
Number of pages4
StatePublished - Dec 1 2005
Event2005 IEEE International SOC Conference - Herndon, VA, United States
Duration: Sep 25 2005Sep 28 2005

Publication series

NameProceedings - IEEE International SOC Conference

Other

Other2005 IEEE International SOC Conference
CountryUnited States
CityHerndon, VA
Period9/25/059/28/05

Fingerprint

Energy conservation
Linear programming
Electric potential
Energy utilization
Voltage scaling

All Science Journal Classification (ASJC) codes

  • Engineering(all)

Cite this

Narayanan, S. H. K., Ozturk, O., Kandemir, M., & Karakoy, M. (2005). Workload clustering for increasing energy savings on embedded MPSoCs. In D. Ha, R. Krishnamurthy, S. Kim, & A. Marshall (Eds.), Proceedings - IEEE International SOC Conference, 2005 SOCC (pp. 157-160). [TA2.1] (Proceedings - IEEE International SOC Conference).
Narayanan, S. H.K. ; Ozturk, O. ; Kandemir, Mahmut ; Karakoy, M. / Workload clustering for increasing energy savings on embedded MPSoCs. Proceedings - IEEE International SOC Conference, 2005 SOCC. editor / D. Ha ; R. Krishnamurthy ; S. Kim ; A. Marshall. 2005. pp. 157-160 (Proceedings - IEEE International SOC Conference).
@inproceedings{934497002b5b41428349280dbcf3d673,
title = "Workload clustering for increasing energy savings on embedded MPSoCs",
abstract = "Voltage/frequency scaling and processor low-power modes (i. e., processor shut-down) are two important mechanisms used for reducing energy consumption in embedded MPSoCs. While a unified scheme that combines these two mechanisms can achieve significant savings in some cases, such an approach is limited by the code parallelization strategy employed. In this paper, we propose a novel, integer linear programming (ILP) based workload clustering strategy across parallel processors, oriented towards maximizing the number of idle processors without impacting original execution times. These idle processors can then be switched to a low power mode to maximize energy savings, whereas the remaining ones can make use of voltage/frequency scaling. In order to check whether this approach brings any energy benefits over the pure voltage scaling based, pure processor shut-down based, or a simple unified scheme, we implemented four different approaches and tested them using a set of eight array/loop-intensive embedded applications. Our simulation-based analysis reveals that the proposed ILP based approach (1) is very effective in reducing the energy consumptions of the applications tested and (2) generates much better energy savings than all the alternate schemes tested (including a unified scheme that combines voltage/frequency scaling and processor shutdown).",
author = "Narayanan, {S. H.K.} and O. Ozturk and Mahmut Kandemir and M. Karakoy",
year = "2005",
month = "12",
day = "1",
language = "English (US)",
isbn = "0780392647",
series = "Proceedings - IEEE International SOC Conference",
pages = "157--160",
editor = "D. Ha and R. Krishnamurthy and S. Kim and A. Marshall",
booktitle = "Proceedings - IEEE International SOC Conference, 2005 SOCC",

}

Narayanan, SHK, Ozturk, O, Kandemir, M & Karakoy, M 2005, Workload clustering for increasing energy savings on embedded MPSoCs. in D Ha, R Krishnamurthy, S Kim & A Marshall (eds), Proceedings - IEEE International SOC Conference, 2005 SOCC., TA2.1, Proceedings - IEEE International SOC Conference, pp. 157-160, 2005 IEEE International SOC Conference, Herndon, VA, United States, 9/25/05.

Workload clustering for increasing energy savings on embedded MPSoCs. / Narayanan, S. H.K.; Ozturk, O.; Kandemir, Mahmut; Karakoy, M.

Proceedings - IEEE International SOC Conference, 2005 SOCC. ed. / D. Ha; R. Krishnamurthy; S. Kim; A. Marshall. 2005. p. 157-160 TA2.1 (Proceedings - IEEE International SOC Conference).

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

TY - GEN

T1 - Workload clustering for increasing energy savings on embedded MPSoCs

AU - Narayanan, S. H.K.

AU - Ozturk, O.

AU - Kandemir, Mahmut

AU - Karakoy, M.

PY - 2005/12/1

Y1 - 2005/12/1

N2 - Voltage/frequency scaling and processor low-power modes (i. e., processor shut-down) are two important mechanisms used for reducing energy consumption in embedded MPSoCs. While a unified scheme that combines these two mechanisms can achieve significant savings in some cases, such an approach is limited by the code parallelization strategy employed. In this paper, we propose a novel, integer linear programming (ILP) based workload clustering strategy across parallel processors, oriented towards maximizing the number of idle processors without impacting original execution times. These idle processors can then be switched to a low power mode to maximize energy savings, whereas the remaining ones can make use of voltage/frequency scaling. In order to check whether this approach brings any energy benefits over the pure voltage scaling based, pure processor shut-down based, or a simple unified scheme, we implemented four different approaches and tested them using a set of eight array/loop-intensive embedded applications. Our simulation-based analysis reveals that the proposed ILP based approach (1) is very effective in reducing the energy consumptions of the applications tested and (2) generates much better energy savings than all the alternate schemes tested (including a unified scheme that combines voltage/frequency scaling and processor shutdown).

AB - Voltage/frequency scaling and processor low-power modes (i. e., processor shut-down) are two important mechanisms used for reducing energy consumption in embedded MPSoCs. While a unified scheme that combines these two mechanisms can achieve significant savings in some cases, such an approach is limited by the code parallelization strategy employed. In this paper, we propose a novel, integer linear programming (ILP) based workload clustering strategy across parallel processors, oriented towards maximizing the number of idle processors without impacting original execution times. These idle processors can then be switched to a low power mode to maximize energy savings, whereas the remaining ones can make use of voltage/frequency scaling. In order to check whether this approach brings any energy benefits over the pure voltage scaling based, pure processor shut-down based, or a simple unified scheme, we implemented four different approaches and tested them using a set of eight array/loop-intensive embedded applications. Our simulation-based analysis reveals that the proposed ILP based approach (1) is very effective in reducing the energy consumptions of the applications tested and (2) generates much better energy savings than all the alternate schemes tested (including a unified scheme that combines voltage/frequency scaling and processor shutdown).

UR - http://www.scopus.com/inward/record.url?scp=30844446708&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=30844446708&partnerID=8YFLogxK

M3 - Conference contribution

SN - 0780392647

SN - 9780780392649

T3 - Proceedings - IEEE International SOC Conference

SP - 157

EP - 160

BT - Proceedings - IEEE International SOC Conference, 2005 SOCC

A2 - Ha, D.

A2 - Krishnamurthy, R.

A2 - Kim, S.

A2 - Marshall, A.

ER -

Narayanan SHK, Ozturk O, Kandemir M, Karakoy M. Workload clustering for increasing energy savings on embedded MPSoCs. In Ha D, Krishnamurthy R, Kim S, Marshall A, editors, Proceedings - IEEE International SOC Conference, 2005 SOCC. 2005. p. 157-160. TA2.1. (Proceedings - IEEE International SOC Conference).