STEAM

A smart temperature and energy aware multicore controller

Vinay Hanumaiah, Digant Desai, Benjamin Gaudette, Carole-Jean Wu, Sarma Vrudhula

Research output: Contribution to journalArticle

16 Citations (Scopus)

Abstract

Recent empirical studies have shown thatmulticore scaling is fast becoming power limited, and consequently, an increasing fraction of a multicore processor has to be under clocked or powered off. Therefore, in addition to fundamental innovations in architecture, compilers and parallelization of application programs, there is a need to develop practical and effective dynamic energy management (DEM) techniques for multicore processors. ExistingDEMtechniques mainly target reducing processor power consumption and temperature, and only few of them have addressed improving energy efficiency for multicore systems.With energy efficiency taking a center stage in all aspects of computing, the focus of the DEM needs to be on finding practical methods to maximize processor efficiency. Towards this, this article presents STEAM - an optimal closed-loop DEM controller designed for multicore processors. The objective is to maximize energy efficiency by dynamic voltage and frequency scaling (DVFS). Energy efficiency is defined as the ratio of performance to power consumption or performance-per-watt (PPW). This is the same as the number of instructions executed per Joule. The PPW metric is actually replaced by PαPW (performanceα-per-Watt), which allows for controlling the importance of performance versus power consumption by varying α. The proposed controller was implemented on a Linux system and tested with the Intel Sandy Bridge processor. There are three power management schemes called governors, available with Intel platforms. They are referred to as (1) Powersave (lowest power consumption), (2) Performance (achieves highest performance), and (3) Ondemand. Our simple and lightweight controller when executing SPEC CPU2006, PARSEC, and MiBench benchmarks have achieved an average of 18% improvement in energy efficiency (MIPS/Watt) over these ACPI policies. Moreover, STEAM also demonstrated an excellent prediction of core temperatures and power consumption, and the ability to control the core temperatures wihin 3°C of the specified maximum. Finally, the overhead of the STEAM implementation (in terms of CPU resources) is less than 0.25%. The entire implementation is self-contained and can be installed on any processor with very little prior knowledge of the processor.

Original languageEnglish (US)
Article number151
JournalACM Transactions on Embedded Computing Systems
Volume13
DOIs
StatePublished - Oct 6 2014

Fingerprint

Energy efficiency
Electric power utilization
Energy management
Controllers
Temperature
Governors
Application programs
Program processors
Innovation

Keywords

  • Closed-loop controller
  • Dynamic energy management
  • Dynamic thermal management
  • Dynamic voltage and frequency scaling
  • Energy efficiency
  • Kalman filter
  • Leakage power dependence on temperature
  • Least-squares estimation
  • Multicore
  • Performance optimal
  • Power and temperature modeling

ASJC Scopus subject areas

  • Hardware and Architecture
  • Software

Cite this

STEAM : A smart temperature and energy aware multicore controller. / Hanumaiah, Vinay; Desai, Digant; Gaudette, Benjamin; Wu, Carole-Jean; Vrudhula, Sarma.

In: ACM Transactions on Embedded Computing Systems, Vol. 13, 151, 06.10.2014.

Research output: Contribution to journalArticle

@article{fbfded55cf0442d2b92b8e39ca98e3fb,
title = "STEAM: A smart temperature and energy aware multicore controller",
abstract = "Recent empirical studies have shown thatmulticore scaling is fast becoming power limited, and consequently, an increasing fraction of a multicore processor has to be under clocked or powered off. Therefore, in addition to fundamental innovations in architecture, compilers and parallelization of application programs, there is a need to develop practical and effective dynamic energy management (DEM) techniques for multicore processors. ExistingDEMtechniques mainly target reducing processor power consumption and temperature, and only few of them have addressed improving energy efficiency for multicore systems.With energy efficiency taking a center stage in all aspects of computing, the focus of the DEM needs to be on finding practical methods to maximize processor efficiency. Towards this, this article presents STEAM - an optimal closed-loop DEM controller designed for multicore processors. The objective is to maximize energy efficiency by dynamic voltage and frequency scaling (DVFS). Energy efficiency is defined as the ratio of performance to power consumption or performance-per-watt (PPW). This is the same as the number of instructions executed per Joule. The PPW metric is actually replaced by PαPW (performanceα-per-Watt), which allows for controlling the importance of performance versus power consumption by varying α. The proposed controller was implemented on a Linux system and tested with the Intel Sandy Bridge processor. There are three power management schemes called governors, available with Intel platforms. They are referred to as (1) Powersave (lowest power consumption), (2) Performance (achieves highest performance), and (3) Ondemand. Our simple and lightweight controller when executing SPEC CPU2006, PARSEC, and MiBench benchmarks have achieved an average of 18{\%} improvement in energy efficiency (MIPS/Watt) over these ACPI policies. Moreover, STEAM also demonstrated an excellent prediction of core temperatures and power consumption, and the ability to control the core temperatures wihin 3°C of the specified maximum. Finally, the overhead of the STEAM implementation (in terms of CPU resources) is less than 0.25{\%}. The entire implementation is self-contained and can be installed on any processor with very little prior knowledge of the processor.",
keywords = "Closed-loop controller, Dynamic energy management, Dynamic thermal management, Dynamic voltage and frequency scaling, Energy efficiency, Kalman filter, Leakage power dependence on temperature, Least-squares estimation, Multicore, Performance optimal, Power and temperature modeling",
author = "Vinay Hanumaiah and Digant Desai and Benjamin Gaudette and Carole-Jean Wu and Sarma Vrudhula",
year = "2014",
month = "10",
day = "6",
doi = "10.1145/2661430",
language = "English (US)",
volume = "13",
journal = "ACM Transactions on Embedded Computing Systems",
issn = "1539-9087",
publisher = "Association for Computing Machinery (ACM)",

}

TY - JOUR

T1 - STEAM

T2 - A smart temperature and energy aware multicore controller

AU - Hanumaiah, Vinay

AU - Desai, Digant

AU - Gaudette, Benjamin

AU - Wu, Carole-Jean

AU - Vrudhula, Sarma

PY - 2014/10/6

Y1 - 2014/10/6

N2 - Recent empirical studies have shown thatmulticore scaling is fast becoming power limited, and consequently, an increasing fraction of a multicore processor has to be under clocked or powered off. Therefore, in addition to fundamental innovations in architecture, compilers and parallelization of application programs, there is a need to develop practical and effective dynamic energy management (DEM) techniques for multicore processors. ExistingDEMtechniques mainly target reducing processor power consumption and temperature, and only few of them have addressed improving energy efficiency for multicore systems.With energy efficiency taking a center stage in all aspects of computing, the focus of the DEM needs to be on finding practical methods to maximize processor efficiency. Towards this, this article presents STEAM - an optimal closed-loop DEM controller designed for multicore processors. The objective is to maximize energy efficiency by dynamic voltage and frequency scaling (DVFS). Energy efficiency is defined as the ratio of performance to power consumption or performance-per-watt (PPW). This is the same as the number of instructions executed per Joule. The PPW metric is actually replaced by PαPW (performanceα-per-Watt), which allows for controlling the importance of performance versus power consumption by varying α. The proposed controller was implemented on a Linux system and tested with the Intel Sandy Bridge processor. There are three power management schemes called governors, available with Intel platforms. They are referred to as (1) Powersave (lowest power consumption), (2) Performance (achieves highest performance), and (3) Ondemand. Our simple and lightweight controller when executing SPEC CPU2006, PARSEC, and MiBench benchmarks have achieved an average of 18% improvement in energy efficiency (MIPS/Watt) over these ACPI policies. Moreover, STEAM also demonstrated an excellent prediction of core temperatures and power consumption, and the ability to control the core temperatures wihin 3°C of the specified maximum. Finally, the overhead of the STEAM implementation (in terms of CPU resources) is less than 0.25%. The entire implementation is self-contained and can be installed on any processor with very little prior knowledge of the processor.

AB - Recent empirical studies have shown thatmulticore scaling is fast becoming power limited, and consequently, an increasing fraction of a multicore processor has to be under clocked or powered off. Therefore, in addition to fundamental innovations in architecture, compilers and parallelization of application programs, there is a need to develop practical and effective dynamic energy management (DEM) techniques for multicore processors. ExistingDEMtechniques mainly target reducing processor power consumption and temperature, and only few of them have addressed improving energy efficiency for multicore systems.With energy efficiency taking a center stage in all aspects of computing, the focus of the DEM needs to be on finding practical methods to maximize processor efficiency. Towards this, this article presents STEAM - an optimal closed-loop DEM controller designed for multicore processors. The objective is to maximize energy efficiency by dynamic voltage and frequency scaling (DVFS). Energy efficiency is defined as the ratio of performance to power consumption or performance-per-watt (PPW). This is the same as the number of instructions executed per Joule. The PPW metric is actually replaced by PαPW (performanceα-per-Watt), which allows for controlling the importance of performance versus power consumption by varying α. The proposed controller was implemented on a Linux system and tested with the Intel Sandy Bridge processor. There are three power management schemes called governors, available with Intel platforms. They are referred to as (1) Powersave (lowest power consumption), (2) Performance (achieves highest performance), and (3) Ondemand. Our simple and lightweight controller when executing SPEC CPU2006, PARSEC, and MiBench benchmarks have achieved an average of 18% improvement in energy efficiency (MIPS/Watt) over these ACPI policies. Moreover, STEAM also demonstrated an excellent prediction of core temperatures and power consumption, and the ability to control the core temperatures wihin 3°C of the specified maximum. Finally, the overhead of the STEAM implementation (in terms of CPU resources) is less than 0.25%. The entire implementation is self-contained and can be installed on any processor with very little prior knowledge of the processor.

KW - Closed-loop controller

KW - Dynamic energy management

KW - Dynamic thermal management

KW - Dynamic voltage and frequency scaling

KW - Energy efficiency

KW - Kalman filter

KW - Leakage power dependence on temperature

KW - Least-squares estimation

KW - Multicore

KW - Performance optimal

KW - Power and temperature modeling

UR - http://www.scopus.com/inward/record.url?scp=84908221896&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84908221896&partnerID=8YFLogxK

U2 - 10.1145/2661430

DO - 10.1145/2661430

M3 - Article

VL - 13

JO - ACM Transactions on Embedded Computing Systems

JF - ACM Transactions on Embedded Computing Systems

SN - 1539-9087

M1 - 151

ER -