- ホーム
- > 洋書
- > 英文書
- > Computer / General
Full Description
WelcometotheproceedingsofthePower-AwareComputerSystems(PACS2002) workshopheld in conjunction with the 8th InternationalSymposium on High PerformanceComputerArchitecture(HPCA-8). Improvementsincomputers- temperformancehavebeenaccompaniedbyanalarmingincreaseinpowerand energydissipation,leading tohigher costandlowerreliabilityinallcomputer systemsmarketsegments. Thehigherpower/energydissipationhasalsosign- icantlyreducedbatterylife inportablesystems. Whilecircuit-leveltechniques continuetoreducepowerandenergy,alllevelsofcomputersystemsarebeing usedtoaddresspowerandenergyissues. PACS2002wasthesecondworkshopin itsseriestoaddresspower-/energy-awarenessatalllevelsofcomputersystems andbroughttogetherexpertsfromacademiaandindustry. Theseproceedingsincluderesearchpapersspanningawidespectrumof- eas in power-aware systems. We have grouped the papers into the following categories:(1)power-awarearchitectureandmicroarchitecture,(2)power-aware real-time systems, (3) power modeling and monitoring, and (4) power-aware operatingsystemsandcompilers. The?rstgroupofpapersproposepower-awaretechniquesfortheprocessor pipeline using adaptiveresizing of power-hungrymicroarchitecturalstructures andclockgating,andpower-awarecachedesignbyavoidingtagchecksin- riodswhenthetagshavenotchanged.
Thisgroupalsoincludesideastoadapt energyandperformancedynamicallybydetectingregionsofapplicationatr- timewherethesupplyvoltagemaybescaledtoreducepowerwithabounded decrease in performance. Lastly, a paper on multiprocessor designs trades o? computingcapacityandfunctionalityforimprovedenergypercyclebysched- ing simple tasks on low-end and low-energy processorsand complex tasks on high-endprocessors. Thesecondgroupofpaperstargetreal-timesystemsincludingideasonal- complexityheuristicwhichschedulesreal-timetaskssuchthatnotaskmissesits deadlineandthetotalenergysavingsaremaximized. Theotherpapersinthis group(1)tunethesystem-levelparallelismtothecurrent-levelofpower/energy availabilityandoptimizethesystempowerutilization,and(2)performadaptive texturemappinginreal-time3Dgraphicssystemsbasedonamodelofhuman visualperceptiontoachievesigni?cantpowersavingswithoutnoticeableimage qualitydegradation. Thethirdgroupofpapersfocusonpowermodelingandmonitoringincluding statisticalpro?lingtodetectsoftwarehotspotsofpower,andusingPetriNetsto modelDRAMpowerpolicies. Thisgroupalsoincludesasimulatorforevaluating theperformanceandpowerofdynamicvoltagescalingalgorithms.
Thelast groupconcentratesonOS and compilersfor lowpower. The ?rst paperproposesapplication-issueddirectivestosetthepowermodesindevices suchasadiskdrive. Thesecondpaperproposespoliciesforcluster-widepower VI Preface management. Thepoliciesemploycombinationsofdynamicvoltagescalingand turningonando?toreduceoverallclusterpower. PACS2002wasahighlysuccessfulforumduetothehigh-qualitysubmissions, theenormouse?ortsoftheprogramcommitteeandthekeynotespeaker,andthe attendees. WewouldliketothankRonnyRonenforanexcellentkeynotespeech, showingthetechnologicalscalingtrendsandtheirimpactonenergy/powerc- sumption in general-purposemicroprocessors,and pinpointing recentmicro- chitecturalstrategiestoachievemorepower-e?cientmicroprocessors. Wewould like to also thank Antonio Gonzalez, Andreas Moshovos,John Kalamatianos, andothermembersoftheHPCA-8organizingcommitteewhohelpedarrange forlocalaccomodationandpublicizetheworkshop. February2002 BabakFalsa?andT. N. Vijaykumar PACS2002 Program Committee BabakFalsa?,CarnegieMellonUniversity(co-chair) T. N.
Vijaykumar,PurdueUniversity(co-chair) DaveAlbonesi,UniversityofRochester KrsteAsanovic,MassachusettsInstituteofTechnology IrisBahar,BrownUniversity LucaBenini,UniversityofBologna DougCarmean,Intel Yuen Chan,IBM KeithFarkas,CompaqWRL MaryJaneIrwin,PennsylvaniaStateUniversity StefanosKaxiras,AgereSystems PeterKogge,UniversityofNotreDame UliKremer,RutgersUniversity AlvinLebeck,DukeUniversity AndreasMoshovos,UniversityofToronto RajRajkumar,CarnegieMellonUniversity KaushikRoy,PurdueUniversity Table of Contents Power-Aware Architecture/Microarchitecture Early-StageDe?nitionofLPX:ALowPowerIssue-ExecuteProcessor ...1 P. Bose,D. Brooks,A. Buyuktosunoglu,P. Cook,K. Das,P. Emma, M. Gschwind,H. Jacobson,T. Karkhanis,P. Kudva,S. Schuster, J. Smith,V. Srinivasan,V. Zyuban,D. Albonesi,andS. Dwarkadas DynamicTag-CheckOmission: ALowPowerInstructionCacheArchitecture ExploitingExecutionFootprints ...18 KojiInoue,VasilyMoshnyaga,andKazuakiMurakami AHardwareArchitecture forDynamicPerformanceandEnergyAdaptation...33 PhillipStanley-Marbell,MichaelS. Hsiao,andUlrichKremer Multi-processorComputerSystemHavingLowPowerConsumption ...53 C. MichaelOlsenandL.
AlexMorrow Power-Aware Real-TimeSystems AnIntegratedHeuristicApproach toPower-AwareReal-TimeScheduling...68 PedroMejia,EugeneLevner,andDanielMoss'e Power-AwareTaskMotionforEnhancingDynamicRange ofEmbeddedSystemswithRenewableEnergySources ...84 JinfengLiu,PaiH. Chou,andNaderBagherzadeh ALow-PowerContent-AdaptiveTextureMappingArchitecture forReal-Time3DGraphics ...99 JeongseonEuh,JeevanChittamuru,andWayneBurleson Power Modelingand Monitoring Energy-DrivenStatisticalSampling:DetectingSoftwareHotspots ...110 FayChang,KeithI. Farkas,andParthasarathyRanganathan ModelingofDRAMPowerControlPolicies UsingDeterministicandStochasticPetriNets ...130 XiaoboFan,CarlaS. Ellis,andAlvinR. Lebeck SimDVS:AnIntegratedSimulationEnvironment forPerformanceEvaluationofDynamicVoltageScalingAlgorithms ...
Contents
Power-Aware Architecture/Microarchitecture.- Early-Stage Definition of LPX: A Low Power Issue-Execute Processor.- Dynamic Tag-Check Omission: A Low Power Instruction Cache Architecture Exploiting Execution Footprints.- A Hardware Architecture for Dynamic Performance and Energy Adaptation.- Multi-Processor Computer System Having Low Power Consumption.- Power-Aware Real-Time Systems.- An Integrated Heuristic Approach to Power-Aware Real-Time Scheduling.- Power-Aware Task Motion for Enhancing Dynamic Range of Embedded Systems with Renewable Energy Sources.- A Low-Power Content-Adaptive Texture Mapping Architecture for Real-Time 3D Graphics.- Power Modeling and Monitoring.- Energy-Driven Statistical Sampling: Detecting Software Hotspots.- Modeling of DRAM Power Control Policies Using Deterministic and Stochastic Petri Nets.- SimDVS: An Integrated Simulation Environment for Performance Evaluation of Dynamic Voltage Scaling Algorithms.- Power-Aware OS and Compilers.- Application-Supported Device Management for Energy and Performance.- Energy-Efficient Server Clusters.- Single Region vs. Multiple Regions: A Comparison of Different Compiler-Directed Dynamic Voltage Scheduling Approaches.
