Application-specific shared last-level cache optimization for low-power embedded systems

Huatao Zhao, Jiongyao Ye, Xian Su, Takahiro Watanabe

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

Abstract

Modern embedded systems favor the chip multiprocessor frame to achieve higher performance, but they are restricted by the inefficient cache hierarchies. Typically, the accessing interference and improper allocation in last-level cache (LLC) shared by multiprocessors cause significant energy consumption and performance depression. In this paper, we propose a configurable and partitioned cache hierarchy where an energy-efficient runtime mechanism can well manage the shared LLC to meet application programs. This mechanism utilizes the repeated behaviors in hot subroutines of application and selects the proper partition intervals. Then, a low-power metric based configurable scheme is employed to explore the optimal allocation of given cache resources. Thus, we can provide each core with the optimal allocation information to dynamically partition the shared LLC during runtime. Experimental results for a quad-core system using the SPEC2006 benchmarks show that the cache access energy can be reduced by on average 32.5 percent compared to the equal partition scheme only with 1.3 percent performance off.

Original languageEnglish
Title of host publicationConference Proceedings - 13th IEEE International NEW Circuits and Systems Conference, NEWCAS 2015
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Print)9781479988938
DOIs
Publication statusPublished - 2015 Aug 6
Event13th IEEE International NEW Circuits and Systems Conference, NEWCAS 2015 - Grenoble, France
Duration: 2015 Jun 72015 Jun 10

Other

Other13th IEEE International NEW Circuits and Systems Conference, NEWCAS 2015
CountryFrance
CityGrenoble
Period15/6/715/6/10

Fingerprint

Subroutines
Application programs
Embedded systems
Energy utilization

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Cite this

Zhao, H., Ye, J., Su, X., & Watanabe, T. (2015). Application-specific shared last-level cache optimization for low-power embedded systems. In Conference Proceedings - 13th IEEE International NEW Circuits and Systems Conference, NEWCAS 2015 [7181994] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/NEWCAS.2015.7181994

Application-specific shared last-level cache optimization for low-power embedded systems. / Zhao, Huatao; Ye, Jiongyao; Su, Xian; Watanabe, Takahiro.

Conference Proceedings - 13th IEEE International NEW Circuits and Systems Conference, NEWCAS 2015. Institute of Electrical and Electronics Engineers Inc., 2015. 7181994.

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

Zhao, H, Ye, J, Su, X & Watanabe, T 2015, Application-specific shared last-level cache optimization for low-power embedded systems. in Conference Proceedings - 13th IEEE International NEW Circuits and Systems Conference, NEWCAS 2015., 7181994, Institute of Electrical and Electronics Engineers Inc., 13th IEEE International NEW Circuits and Systems Conference, NEWCAS 2015, Grenoble, France, 15/6/7. https://doi.org/10.1109/NEWCAS.2015.7181994
Zhao H, Ye J, Su X, Watanabe T. Application-specific shared last-level cache optimization for low-power embedded systems. In Conference Proceedings - 13th IEEE International NEW Circuits and Systems Conference, NEWCAS 2015. Institute of Electrical and Electronics Engineers Inc. 2015. 7181994 https://doi.org/10.1109/NEWCAS.2015.7181994
Zhao, Huatao ; Ye, Jiongyao ; Su, Xian ; Watanabe, Takahiro. / Application-specific shared last-level cache optimization for low-power embedded systems. Conference Proceedings - 13th IEEE International NEW Circuits and Systems Conference, NEWCAS 2015. Institute of Electrical and Electronics Engineers Inc., 2015.
@inproceedings{c9baec1152154c1ca907445d0ebfea31,
title = "Application-specific shared last-level cache optimization for low-power embedded systems",
abstract = "Modern embedded systems favor the chip multiprocessor frame to achieve higher performance, but they are restricted by the inefficient cache hierarchies. Typically, the accessing interference and improper allocation in last-level cache (LLC) shared by multiprocessors cause significant energy consumption and performance depression. In this paper, we propose a configurable and partitioned cache hierarchy where an energy-efficient runtime mechanism can well manage the shared LLC to meet application programs. This mechanism utilizes the repeated behaviors in hot subroutines of application and selects the proper partition intervals. Then, a low-power metric based configurable scheme is employed to explore the optimal allocation of given cache resources. Thus, we can provide each core with the optimal allocation information to dynamically partition the shared LLC during runtime. Experimental results for a quad-core system using the SPEC2006 benchmarks show that the cache access energy can be reduced by on average 32.5 percent compared to the equal partition scheme only with 1.3 percent performance off.",
author = "Huatao Zhao and Jiongyao Ye and Xian Su and Takahiro Watanabe",
year = "2015",
month = "8",
day = "6",
doi = "10.1109/NEWCAS.2015.7181994",
language = "English",
isbn = "9781479988938",
booktitle = "Conference Proceedings - 13th IEEE International NEW Circuits and Systems Conference, NEWCAS 2015",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Application-specific shared last-level cache optimization for low-power embedded systems

AU - Zhao, Huatao

AU - Ye, Jiongyao

AU - Su, Xian

AU - Watanabe, Takahiro

PY - 2015/8/6

Y1 - 2015/8/6

N2 - Modern embedded systems favor the chip multiprocessor frame to achieve higher performance, but they are restricted by the inefficient cache hierarchies. Typically, the accessing interference and improper allocation in last-level cache (LLC) shared by multiprocessors cause significant energy consumption and performance depression. In this paper, we propose a configurable and partitioned cache hierarchy where an energy-efficient runtime mechanism can well manage the shared LLC to meet application programs. This mechanism utilizes the repeated behaviors in hot subroutines of application and selects the proper partition intervals. Then, a low-power metric based configurable scheme is employed to explore the optimal allocation of given cache resources. Thus, we can provide each core with the optimal allocation information to dynamically partition the shared LLC during runtime. Experimental results for a quad-core system using the SPEC2006 benchmarks show that the cache access energy can be reduced by on average 32.5 percent compared to the equal partition scheme only with 1.3 percent performance off.

AB - Modern embedded systems favor the chip multiprocessor frame to achieve higher performance, but they are restricted by the inefficient cache hierarchies. Typically, the accessing interference and improper allocation in last-level cache (LLC) shared by multiprocessors cause significant energy consumption and performance depression. In this paper, we propose a configurable and partitioned cache hierarchy where an energy-efficient runtime mechanism can well manage the shared LLC to meet application programs. This mechanism utilizes the repeated behaviors in hot subroutines of application and selects the proper partition intervals. Then, a low-power metric based configurable scheme is employed to explore the optimal allocation of given cache resources. Thus, we can provide each core with the optimal allocation information to dynamically partition the shared LLC during runtime. Experimental results for a quad-core system using the SPEC2006 benchmarks show that the cache access energy can be reduced by on average 32.5 percent compared to the equal partition scheme only with 1.3 percent performance off.

UR - http://www.scopus.com/inward/record.url?scp=84945144182&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84945144182&partnerID=8YFLogxK

U2 - 10.1109/NEWCAS.2015.7181994

DO - 10.1109/NEWCAS.2015.7181994

M3 - Conference contribution

AN - SCOPUS:84945144182

SN - 9781479988938

BT - Conference Proceedings - 13th IEEE International NEW Circuits and Systems Conference, NEWCAS 2015

PB - Institute of Electrical and Electronics Engineers Inc.

ER -