Performance optimal processor throttling under thermal constraints

Ravishankar Rao, Sarma Vrudhula

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

35 Citations (Scopus)

Abstract

We derive analytically, the performance optimal throttling curve for a processor under thermal constraints for a given task sequence. We found that keeping the chip temperature constant requires an exponential speed curve. Earlier works that propose constant throttling only keep the package/case temperature constant, and are hence suboptimal. We develop high-level thermal and power models that are simple enough for analysis, yet account for important effects like the power-density variation across a chip (hotspots), leakage dependence on temperature (LDT), and differing thermal characteristics of the silicon die and the thermal solution. We use a piecewise-linear approximation for the exponential leakage dependence on temperature, and devise a method to remove the circular dependency between leakage power and temperature. To solve the multi-task speed control problem, we first solve analytically, the single task problem with a constraint on the final package temperature using optimal control theory. We then find the optimum final package temperature of each task by dynamic programming. We compared the total execution time of several randomly generated task sequences using the optimal control policy against a constant speed throttling policy, and found significantly smaller total execution times. We compared the thermal profiles predicted by the proposed high-level thermal model to that of the Hotspot thermal model, and found them to be in good agreement.

Original languageEnglish (US)
Title of host publicationCASES'07: Proceedings of the 2007 International Conference on Compilers, Architecture, and Synthesis for Embedded Systems
Pages257-266
Number of pages10
DOIs
StatePublished - 2007
EventCASES'07: 2007 International Conference on Compilers, Architecture, and Synthesis for Embedded Systems - Salzburg, Austria
Duration: Sep 30 2007Oct 3 2007

Other

OtherCASES'07: 2007 International Conference on Compilers, Architecture, and Synthesis for Embedded Systems
CountryAustria
CitySalzburg
Period9/30/0710/3/07

Fingerprint

Temperature
Speed control
Hot Temperature
Control theory
Dynamic programming
Silicon

Keywords

  • Leakage dependence on temperature
  • Power
  • Thermal management
  • Thermal model
  • Throttling

ASJC Scopus subject areas

  • Hardware and Architecture
  • Software

Cite this

Rao, R., & Vrudhula, S. (2007). Performance optimal processor throttling under thermal constraints. In CASES'07: Proceedings of the 2007 International Conference on Compilers, Architecture, and Synthesis for Embedded Systems (pp. 257-266) https://doi.org/10.1145/1289881.1289925

Performance optimal processor throttling under thermal constraints. / Rao, Ravishankar; Vrudhula, Sarma.

CASES'07: Proceedings of the 2007 International Conference on Compilers, Architecture, and Synthesis for Embedded Systems. 2007. p. 257-266.

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

Rao, R & Vrudhula, S 2007, Performance optimal processor throttling under thermal constraints. in CASES'07: Proceedings of the 2007 International Conference on Compilers, Architecture, and Synthesis for Embedded Systems. pp. 257-266, CASES'07: 2007 International Conference on Compilers, Architecture, and Synthesis for Embedded Systems, Salzburg, Austria, 9/30/07. https://doi.org/10.1145/1289881.1289925
Rao R, Vrudhula S. Performance optimal processor throttling under thermal constraints. In CASES'07: Proceedings of the 2007 International Conference on Compilers, Architecture, and Synthesis for Embedded Systems. 2007. p. 257-266 https://doi.org/10.1145/1289881.1289925
Rao, Ravishankar ; Vrudhula, Sarma. / Performance optimal processor throttling under thermal constraints. CASES'07: Proceedings of the 2007 International Conference on Compilers, Architecture, and Synthesis for Embedded Systems. 2007. pp. 257-266
@inproceedings{dcdbb3029d70434b830aaef0e83849ec,
title = "Performance optimal processor throttling under thermal constraints",
abstract = "We derive analytically, the performance optimal throttling curve for a processor under thermal constraints for a given task sequence. We found that keeping the chip temperature constant requires an exponential speed curve. Earlier works that propose constant throttling only keep the package/case temperature constant, and are hence suboptimal. We develop high-level thermal and power models that are simple enough for analysis, yet account for important effects like the power-density variation across a chip (hotspots), leakage dependence on temperature (LDT), and differing thermal characteristics of the silicon die and the thermal solution. We use a piecewise-linear approximation for the exponential leakage dependence on temperature, and devise a method to remove the circular dependency between leakage power and temperature. To solve the multi-task speed control problem, we first solve analytically, the single task problem with a constraint on the final package temperature using optimal control theory. We then find the optimum final package temperature of each task by dynamic programming. We compared the total execution time of several randomly generated task sequences using the optimal control policy against a constant speed throttling policy, and found significantly smaller total execution times. We compared the thermal profiles predicted by the proposed high-level thermal model to that of the Hotspot thermal model, and found them to be in good agreement.",
keywords = "Leakage dependence on temperature, Power, Thermal management, Thermal model, Throttling",
author = "Ravishankar Rao and Sarma Vrudhula",
year = "2007",
doi = "10.1145/1289881.1289925",
language = "English (US)",
isbn = "9781595938268",
pages = "257--266",
booktitle = "CASES'07: Proceedings of the 2007 International Conference on Compilers, Architecture, and Synthesis for Embedded Systems",

}

TY - GEN

T1 - Performance optimal processor throttling under thermal constraints

AU - Rao, Ravishankar

AU - Vrudhula, Sarma

PY - 2007

Y1 - 2007

N2 - We derive analytically, the performance optimal throttling curve for a processor under thermal constraints for a given task sequence. We found that keeping the chip temperature constant requires an exponential speed curve. Earlier works that propose constant throttling only keep the package/case temperature constant, and are hence suboptimal. We develop high-level thermal and power models that are simple enough for analysis, yet account for important effects like the power-density variation across a chip (hotspots), leakage dependence on temperature (LDT), and differing thermal characteristics of the silicon die and the thermal solution. We use a piecewise-linear approximation for the exponential leakage dependence on temperature, and devise a method to remove the circular dependency between leakage power and temperature. To solve the multi-task speed control problem, we first solve analytically, the single task problem with a constraint on the final package temperature using optimal control theory. We then find the optimum final package temperature of each task by dynamic programming. We compared the total execution time of several randomly generated task sequences using the optimal control policy against a constant speed throttling policy, and found significantly smaller total execution times. We compared the thermal profiles predicted by the proposed high-level thermal model to that of the Hotspot thermal model, and found them to be in good agreement.

AB - We derive analytically, the performance optimal throttling curve for a processor under thermal constraints for a given task sequence. We found that keeping the chip temperature constant requires an exponential speed curve. Earlier works that propose constant throttling only keep the package/case temperature constant, and are hence suboptimal. We develop high-level thermal and power models that are simple enough for analysis, yet account for important effects like the power-density variation across a chip (hotspots), leakage dependence on temperature (LDT), and differing thermal characteristics of the silicon die and the thermal solution. We use a piecewise-linear approximation for the exponential leakage dependence on temperature, and devise a method to remove the circular dependency between leakage power and temperature. To solve the multi-task speed control problem, we first solve analytically, the single task problem with a constraint on the final package temperature using optimal control theory. We then find the optimum final package temperature of each task by dynamic programming. We compared the total execution time of several randomly generated task sequences using the optimal control policy against a constant speed throttling policy, and found significantly smaller total execution times. We compared the thermal profiles predicted by the proposed high-level thermal model to that of the Hotspot thermal model, and found them to be in good agreement.

KW - Leakage dependence on temperature

KW - Power

KW - Thermal management

KW - Thermal model

KW - Throttling

UR - http://www.scopus.com/inward/record.url?scp=38849163446&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=38849163446&partnerID=8YFLogxK

U2 - 10.1145/1289881.1289925

DO - 10.1145/1289881.1289925

M3 - Conference contribution

AN - SCOPUS:38849163446

SN - 9781595938268

SP - 257

EP - 266

BT - CASES'07: Proceedings of the 2007 International Conference on Compilers, Architecture, and Synthesis for Embedded Systems

ER -