Green computing: power optimisation of VFI-based real-time multiprocessor dataflow applications (extended version)

Abstract

Execution time is no longer the only performance metric for computer systems. In fact, a trend is emerging to trade raw performance for energy savings. Techniques like Dynamic Power Management (DPM, switching to low power state) and Dynamic Voltage and Frequency Scaling (DVFS, throttling processor frequency) help modern systems to reduce their power consumption while adhering to performance requirements. To balance flexibility and design complexity, the concept of Voltage and Frequency Islands (VFIs) was recently introduced for power optimisation. It achieves fine-grained system-level power management, by operating all processors in the same VFI at a common frequency/voltage.This paper presents a novel approach to compute a power management strategy combining DPM and DVFS. In our approach, applications (modelled in full synchronous dataflow, SDF) are mapped on heterogeneous multiprocessor platforms (partitioned in voltage and frequency islands). We compute an energy-optimal schedule, meeting minimal throughput requirements. We demonstrate that the combination of DPM and DVFS provides an energy reduction beyond considering DVFS or DMP separately. Moreover, we show that by clustering processors in VFIs, DPM can be combined with any granularity of DVFS. Our approach uses model checking, by encoding the optimisation problem as a query over priced timed automata. The model-checker Uppaal Cora extracts a cost minimal trace, representing a power minimal schedule. We illustrate our approach with several case studies on commercially available hardware.
Original languageUndefined
Place of PublicationEnschede
PublisherCentre for Telematics and Information Technology (CTIT)
Number of pages32
StatePublished - 24 Jun 2015

Publication series

NameCTIT Technical Report Series
PublisherUniversity of Twente, Centre for Telematics and Information Technology (CTIT)
No.TR-CTIT-15-04
ISSN (Print)1381-3625

Fingerprint

Power management
Model checking
Energy conservation
Computer systems
Electric power utilization
Throughput
Hardware
Costs
Voltage scaling
Dynamic frequency scaling

Keywords

  • Throughput
  • Voltage and Frequency Islands
  • IR-96408
  • Transition overheads
  • EC Grant Agreement nr.: FP7/318490
  • Heterogeneous
  • EC Grant Agreement nr.: FP7/2007-2013
  • Data flow
  • Green Computing
  • Dynamic Voltage and Frequency Scaling (DVFS)
  • Dynamic Power Management (DPM)
  • Priced Timed Automata
  • UPPAAL
  • METIS-312652
  • Power Minimisation
  • EWI-26102

Cite this

Ahmad, W., Holzenspies, P. K. F., Stoelinga, M. I. A., & van de Pol, J. C. (2015). Green computing: power optimisation of VFI-based real-time multiprocessor dataflow applications (extended version). (CTIT Technical Report Series; No. TR-CTIT-15-04). Enschede: Centre for Telematics and Information Technology (CTIT).

Ahmad, W.; Holzenspies, P.K.F.; Stoelinga, Mariëlle Ida Antoinette; van de Pol, Jan Cornelis / Green computing: power optimisation of VFI-based real-time multiprocessor dataflow applications (extended version).

Enschede : Centre for Telematics and Information Technology (CTIT), 2015. 32 p. (CTIT Technical Report Series; No. TR-CTIT-15-04).

Research output: ProfessionalReport

@book{24ea7d3e1bf846b4b7ec8ec797e21e6b,
title = "Green computing: power optimisation of VFI-based real-time multiprocessor dataflow applications (extended version)",
abstract = "Execution time is no longer the only performance metric for computer systems. In fact, a trend is emerging to trade raw performance for energy savings. Techniques like Dynamic Power Management (DPM, switching to low power state) and Dynamic Voltage and Frequency Scaling (DVFS, throttling processor frequency) help modern systems to reduce their power consumption while adhering to performance requirements. To balance flexibility and design complexity, the concept of Voltage and Frequency Islands (VFIs) was recently introduced for power optimisation. It achieves fine-grained system-level power management, by operating all processors in the same VFI at a common frequency/voltage.This paper presents a novel approach to compute a power management strategy combining DPM and DVFS. In our approach, applications (modelled in full synchronous dataflow, SDF) are mapped on heterogeneous multiprocessor platforms (partitioned in voltage and frequency islands). We compute an energy-optimal schedule, meeting minimal throughput requirements. We demonstrate that the combination of DPM and DVFS provides an energy reduction beyond considering DVFS or DMP separately. Moreover, we show that by clustering processors in VFIs, DPM can be combined with any granularity of DVFS. Our approach uses model checking, by encoding the optimisation problem as a query over priced timed automata. The model-checker Uppaal Cora extracts a cost minimal trace, representing a power minimal schedule. We illustrate our approach with several case studies on commercially available hardware.",
keywords = "Throughput, Voltage and Frequency Islands, IR-96408, Transition overheads, EC Grant Agreement nr.: FP7/318490, Heterogeneous, EC Grant Agreement nr.: FP7/2007-2013, Data flow, Green Computing, Dynamic Voltage and Frequency Scaling (DVFS), Dynamic Power Management (DPM), Priced Timed Automata, UPPAAL, METIS-312652, Power Minimisation, EWI-26102",
author = "W. Ahmad and P.K.F. Holzenspies and Stoelinga, {Mariëlle Ida Antoinette} and {van de Pol}, {Jan Cornelis}",
note = "Technical Report accompanying DSD 2015 paper.",
year = "2015",
month = "6",
series = "CTIT Technical Report Series",
publisher = "Centre for Telematics and Information Technology (CTIT)",
number = "TR-CTIT-15-04",
address = "Netherlands",

}

Ahmad, W, Holzenspies, PKF, Stoelinga, MIA & van de Pol, JC 2015, Green computing: power optimisation of VFI-based real-time multiprocessor dataflow applications (extended version). CTIT Technical Report Series, no. TR-CTIT-15-04, Centre for Telematics and Information Technology (CTIT), Enschede.

Green computing: power optimisation of VFI-based real-time multiprocessor dataflow applications (extended version). / Ahmad, W.; Holzenspies, P.K.F.; Stoelinga, Mariëlle Ida Antoinette; van de Pol, Jan Cornelis.

Enschede : Centre for Telematics and Information Technology (CTIT), 2015. 32 p. (CTIT Technical Report Series; No. TR-CTIT-15-04).

Research output: ProfessionalReport

TY - BOOK

T1 - Green computing: power optimisation of VFI-based real-time multiprocessor dataflow applications (extended version)

AU - Ahmad,W.

AU - Holzenspies,P.K.F.

AU - Stoelinga,Mariëlle Ida Antoinette

AU - van de Pol,Jan Cornelis

N1 - Technical Report accompanying DSD 2015 paper.

PY - 2015/6/24

Y1 - 2015/6/24

N2 - Execution time is no longer the only performance metric for computer systems. In fact, a trend is emerging to trade raw performance for energy savings. Techniques like Dynamic Power Management (DPM, switching to low power state) and Dynamic Voltage and Frequency Scaling (DVFS, throttling processor frequency) help modern systems to reduce their power consumption while adhering to performance requirements. To balance flexibility and design complexity, the concept of Voltage and Frequency Islands (VFIs) was recently introduced for power optimisation. It achieves fine-grained system-level power management, by operating all processors in the same VFI at a common frequency/voltage.This paper presents a novel approach to compute a power management strategy combining DPM and DVFS. In our approach, applications (modelled in full synchronous dataflow, SDF) are mapped on heterogeneous multiprocessor platforms (partitioned in voltage and frequency islands). We compute an energy-optimal schedule, meeting minimal throughput requirements. We demonstrate that the combination of DPM and DVFS provides an energy reduction beyond considering DVFS or DMP separately. Moreover, we show that by clustering processors in VFIs, DPM can be combined with any granularity of DVFS. Our approach uses model checking, by encoding the optimisation problem as a query over priced timed automata. The model-checker Uppaal Cora extracts a cost minimal trace, representing a power minimal schedule. We illustrate our approach with several case studies on commercially available hardware.

AB - Execution time is no longer the only performance metric for computer systems. In fact, a trend is emerging to trade raw performance for energy savings. Techniques like Dynamic Power Management (DPM, switching to low power state) and Dynamic Voltage and Frequency Scaling (DVFS, throttling processor frequency) help modern systems to reduce their power consumption while adhering to performance requirements. To balance flexibility and design complexity, the concept of Voltage and Frequency Islands (VFIs) was recently introduced for power optimisation. It achieves fine-grained system-level power management, by operating all processors in the same VFI at a common frequency/voltage.This paper presents a novel approach to compute a power management strategy combining DPM and DVFS. In our approach, applications (modelled in full synchronous dataflow, SDF) are mapped on heterogeneous multiprocessor platforms (partitioned in voltage and frequency islands). We compute an energy-optimal schedule, meeting minimal throughput requirements. We demonstrate that the combination of DPM and DVFS provides an energy reduction beyond considering DVFS or DMP separately. Moreover, we show that by clustering processors in VFIs, DPM can be combined with any granularity of DVFS. Our approach uses model checking, by encoding the optimisation problem as a query over priced timed automata. The model-checker Uppaal Cora extracts a cost minimal trace, representing a power minimal schedule. We illustrate our approach with several case studies on commercially available hardware.

KW - Throughput

KW - Voltage and Frequency Islands

KW - IR-96408

KW - Transition overheads

KW - EC Grant Agreement nr.: FP7/318490

KW - Heterogeneous

KW - EC Grant Agreement nr.: FP7/2007-2013

KW - Data flow

KW - Green Computing

KW - Dynamic Voltage and Frequency Scaling (DVFS)

KW - Dynamic Power Management (DPM)

KW - Priced Timed Automata

KW - UPPAAL

KW - METIS-312652

KW - Power Minimisation

KW - EWI-26102

M3 - Report

T3 - CTIT Technical Report Series

BT - Green computing: power optimisation of VFI-based real-time multiprocessor dataflow applications (extended version)

PB - Centre for Telematics and Information Technology (CTIT)

ER -

Ahmad W, Holzenspies PKF, Stoelinga MIA, van de Pol JC. Green computing: power optimisation of VFI-based real-time multiprocessor dataflow applications (extended version). Enschede: Centre for Telematics and Information Technology (CTIT), 2015. 32 p. (CTIT Technical Report Series; TR-CTIT-15-04).