您好,欢迎访问三七文档
TECHNICALREPORT94-8RandomizedOn-lineAlgorithmsforthePageReplicationProblemHisashiKoga(Univ.ofTokyo)April6,1994TDepartmentofInformationScienceFacultyofScience,UniversityofTokyo7{3{1Hongo,Bunkyo-KuTokyo,113JapanTECHNICALREPORT94-8TITLERandomizedOn-lineAlgorithmsforthePageReplicationProblemAUTHORSHisashiKoga(Univ.ofTokyo)KEYWORDSANDPHRASESCompetitiveanalysis,Distributedsharedmemory,Pagereplication,On-linealgo-rithmPotentialfunctionABSTRACTInadistributedsharedmemoryconsistingofmultipleprocessors,eachofwhichhasitsownlocalmemories,eachread-onlypageneedstobelocatedatappropriateprocessorsbyreplicationsoastomakethetotalcostlower.Thepurposeofthepagereplicationproblemistoimplementthislow-costlocating.Thispaperconsiderson-linealgorithmsforthisproblemintermsofcompetitiveness,theratioofthecostofon-linealgorithmstothatoftheo -lineoptimalalgorithms.Especiallyweapplyrandomizedtechniquesdevelopedforthepagemigrational-gorithmstothepagereplicationproblem.Asaresult,we ndrandomizedalgo-rithmsfortreeswhichbreakthedeterministiclowerbound.Moreover,memorylesscoin- ippingalgorithmsareinvestigated.Weprovethatacoin- ippingalgorithmis2-competitivefortrees.Forcircles,wedevelopanewgeneralizedcoin- ippingalgorithmthatis4-competitive.Thisnewalgorithmforcircleshasthelowestcom-petitiveratioamongalreadyknownalgorithms.ANYOTHERIDENTIFYINGINFORMATIONOFTHISREPORTThepreliminaryversionofthispaperwaspresentedatISAAC’93inHongKong.DISTRIBUTIONSTATEMENTFirstissue35copies.SUPPLEMENTARYNOTESREPORTDATEApril6,1994TOTALNO.OFPAGES25WRITTENLANGUAGEEnglishNO.OFREFERENCES13DEPARTMENTOFINFORMATIONSCIENCEFacultyofScience,UniversityofTokyo7-3-1Hongo,Bunkyo-ku,Tokyo113,JapanRandomizedOn-lineAlgorithmsforthePageReplicationProblemHisashiKoga(Univ.ofTokyo)April6,19941IntroductionAcommondesignforasharedmemorymultiprocessorsystemisanetworkofprocessors,eachofwhichhasitsownlocalmemory.Insuchadesign,aprogrammingabstractionofasimpleglobalmemoryissupportedbyavirtualmemorysystemthatdistributesthephysicalpagesamongthelocalmemories.Inadistributedsharedmemorysystem,whenprocessorqwishestoaccessmemoryaddressaofpageb, rstqexamineswhetherthepageiscontainedinitslocalmemory.Ifso,thepageaccessisdonelocallywithnocost.Otherwise,qmustsearchtheprocessorpholdingthepagebandsendamemoryaccessrequesttop.Thentheprocessorprespondstotherequestandthevalueofthelocationaistransmittedbacktoq.Thisactionincursthecostproportionaltothedistancebetweenpandq.Noteeverytimeqaccessesthepage,thispositivecostisrequiredprovidedqdoesnothavethepage.Therefore,ifqrequiresthepageaccesstobsooften,themigration/replicationofafullpageofbmayresultinlowercost,becauseqcanaccessthepageat0cost,afterqhasthepage.Ontheotherhand,movingafullpageincursalargeamountofcommunicationcostproportionaltothepagesize.Inthispaper,wedealwithdistributedsharedmemorysystemswherebroadcast,invalidate,orsnoopingmechanismsarenotallowed.Insuchsystems,awritablepageisrestrictedtohaveonlyonecopytoavoidthedi cultyofmaintainingconsistencyamongmultiplecopies.Therefore,forawritablepageitisimportanttoconsiderthepagemigrationproblem.Thepurposeofthisproblemistodeviseresidencystrategiesthatdecidewhichlocalmemoryshouldhavetheonlycopyofthewritablepageinordertoreducetheamountofcommunicationinprocessingasequenceofpage-accessrequests.Onthecontrary,foraread-onlypage,manycopiesmayexistatthesametime,becausetheconsistencycannotbebroken.Therefore,to ndresidencystrategiesthatdecidewhichsubsetofthelocalmemoriesshouldcontainthecopyofthepageisessential.Thisproblemiscalledthepagereplicationproblem.Inthispaper,wefocusonon-linealgorithmsforthepagereplicationproblem.Analgorithmissaidtobeon-lineifitprocessesarequestbasedonlyonthatrequestandthepastrequests.Toevaluateon-linealgorithms,weusecompetitiveanalysisintroducedin[13].Roughlyspeaking,anon-linealgorithmAisc-competitiveifforanyrequestsequence ,thecostofAtoprocess isnomorethanctimesthecostoftheo -lineoptimalalgorithm.1Thecompetitiveanalysisisinasenseapessimisticapproachbasedonevaluatingtheworst-case,andcanbea ectedbyafewbadexamples.However,byconsideringtheexpectedcostofrandomizedalgorithms,thein uenceofsuchbadexamplesinthecom-petitiveanalysisareweakened.Inthissense,randomizationgivesstrongpowertoon-linealgorithms.Infact,therearemanyon-lineproblemssuchthatrandomizedon-linealgo-rithmsbreakthedeterministiclowerboundofthecompetitiveratio.Thisphenomenonre ectstheabovefact.Forthemigrationproblem,BlackandSleator[4]developed3-competitivedetermin-isticon-linealgorithmsforuniformnetworks(completegraphswitheachedgehavingalengthof1)andtreenetworkswitharbitraryedgelength.Theyalsoshowthatnodeterministicon-linealgorithmcannotbebetterthan3-competitiveevenwhenthegraphconsistsofasingleedge.However,Westbrookshowsthatrandomizationcanbeatthede-terministiclowerbound[14].Hegavea5+p174’2:28-competitiverandomizedalgorithmforuniformnetworkswhenthepagesizefactorislarge.Forgeneralgraphs,hedeviseda2.62-competitiverandomizedalgorithm.Inaddition,heprovedthatasimplecoin- ippingalgorithmis3-competitiveagainstanytopologiesofnetworks.SubsequenttoWestbrook’sresult,C
本文标题:DISTRIBUTION STATEMENT First issue 35 copies. SUPP
链接地址:https://www.777doc.com/doc-3281368 .html