Size: px
Start display at page:

Download ""

Transcription

1 Nearestneighboralgorithmsforloadbalancingin ChengzhongXu parallelcomputers DepartmentofElectricalandComputerEngg. WayneStateUniversity,Detroit,MI48202 BurkhardMonien,ReinhardLuling UniversityofPaderborn,Germany DepartmentofComputerScience DepartmentofComputerScience,TheUniversityofHongKong,HongKong sionsbasedonlocalizedworkloadinformationandmanagesworkloadmigrationswithinits Withnearestneighborloadbalancingalgorithms,aprocessormakesbalancingdeci- Abstract andtheirseveralvariants theaveragedimension-exchange(ade),theoptimally-tuned neighborhood.thispapercomparesacoupleoffairlywell-knownnearestneighboralgorithms,thedimension-exchange(de,forshort)andthediusion(df,forshort)methods dimension-exchange(ode),thelocalaveragediusion(adf)andtheoptimally-tuneddiffusion(odf).themeasuresofinterestaretheireciencyindrivinganyinitialworkload distributiontoauniformdistributionandtheirabilityincontrollingthegrowthofthevarianceamongtheprocessors'workloads.thecomparisonismadewithrespecttobothoneportandall-portcommunicationarchitecturesandinconsiderationofvariousimplementationstrategiesincludingsynchronous/asynchronousinvocationpoliciesandstatic/dynamigorithmleadsitselftobestsuitedforstaticallysynchronousimplementationsofaload thediusionmethodintheone-portcommunicationmodel.inparticular,theodeal- randomworkloadbehaviors.itturnsoutthatthedimension-exchangemethodoutperforms theodfalgorithmperformsbestinthatcase.theunderlyingcommunicationnetworks consideredassumethemostpopulartopologies,themeshandthetorusandtheirspecial diusionmethodisinasynchronousimplementationsintheall-portcommunicationmodel; balancingprocessregardlessofitsunderlyingcommunicationmodels.thestrengthofthe cases:thehypercubeandthek-aryn-cube.

2 Massivelyparallelcomputershavebeenshowntobeveryecientinsolvingproblemsthatcan bepartitionedintotaskswithstaticcomputationandcommunicationpatterns.however,there Introduction communicationpatterns.tosolvethiskindofproblemsecientlyinparallelcomputers,itis existalargeclassofproblemsthathaveunpredictablecomputationalrequirementsorirregular necessarytoperformloadbalancingoperationsatrun-time. totakeplace.everyloadbalancingstrategyhastoresolvetheissuesofwhentoinvokea viewofthesystemandsomenegotiationmechanismforworkloadmigrationsacrossprocessors balancingoperation,whomakesloadbalancingdecisionsaccordingtowhatinformation,and Theexecutionofaloadbalancingprocedurerequiressomemeansofmaintainingaglobal howtomanageworkloadmigrationsbetweenprocessors.combiningdierentanswerstothe aboveyieldsalargespaceofpossibledesignsofloadbalancingalgorithmswithwidelyvarying characteristics.nearestneighboralgorithmsaresuchaclassofmethodsinwhichprocessors workloadtonearestneighbors,thesealgorithmscanbeeasilyscaledtooperateinmassively makedecisionsbasedonlocalinformationinadecentralizedmannerandmanageworkload migrationswithintheimmediateneighborhood[,2,3,4,5].sincetheywouldonlyspreadlocal parallelcomputersofanysize,andwouldtendtopreservethecommunicationlocalityinherent intheunderlyingcomputations.ingeneral,thesealgorithmsareexecutediteratively,withthe expectationthatsuccessiveinvocationsoflocalloadbalancingwouldeventuallybringabouta spectrumofpossibilities,fromloadsharing(noidleprocessorscoexistwithbusyprocessors) globalbalancedstate;hence,theygivetheexibilityofcontrollingthebalancequalityovera uniformdistribution,andhenceateachoperation,needonlybeconcernedwiththedirection totheglobalbalancedstate. loadbalancingmethodsthatarecharacterizedbydierentchoicesofthedirectionofworkload ofworkloadmigrationandtheissueofhowtoapportionexcessworkloads.amongexisting Nearestneighborloadbalancingalgorithmsrelyonsuccessiveapproximationstoaglobal migration[6],weareinterestedinthediusionandthedimension-exchangemethods.these twomethodshavedrawnafairamountofattentioninrecentyears.withthediusionmethod, aheavilyorlightlyloadedprocessorbalancesitsworkloadwithallofitsnearestneighbors thesubsequentpairwisebalancing[8,5,9].thesetwomethodsarecloselyrelated,andthey simultaneouslyinaloadbalancingoperation[7,8].withthethedimension-exchangemethod, oneatatime,andeachtimeanewworkloadindexiscomputed,whichwillbeusedinthe lendthemselvesparticularlywelltoimplementationintwobasiccommunicationarchitectures, aprocessorinneedofloadbalancingbalancesitsworkloadsuccessivelywithitsneighbors theall-portandtheone-portmodels,respectively.theall-portmodelallowsaprocessorto neighboratonetime.bothofthesetwomodelswereassumedinmanyrecentresearcheson exchangemessageswithallitsdirectneighborssimultaneouslyinonecommunicationstep, whiletheone-portmodelrestrictsaprocessortoexchangemessageswithatmostonedirect communicationalgorithms[0,].althoughthelatestdesignsofmessage-passingprocessors tendtosupportall-portsimultaneouscommunications,therestrictiveone-portmodelisstill 2

3 validinexistingrealparallelcomputersystems.sincethecostinsettingupacommunication isxed,thetotaltimespentinsendingdmessagestoddierentports,assumingthebest possibleoverlappingintime,isstilllargelydeterminedbydunlessthemessagesarerather longṫheall-portandone-portmodelsfavorthediusionandthedimension-exchangemethods, usingthediusionmethodcanbecompletedinonecommunicationstepwhilethatusing respectively.inasystemthatsupportsall-portcommunications,aloadbalancingoperation bandwidthisconcerned.anaturalbutinterestingquestioniswhethertheadvantagetranslates hasanadvantageoverthedimension-exchangemethodasfarasexploitingthecommunication thedimension-exchangemethodwouldtakedsteps.itappearsthatthediusionmethod auniformdistribution.thismeasurealoneissucientforthosekindsofproblemsthatneed ofcommunicationstepsrequiredbythealgorithmtodriveaninitialworkloaddistributioninto algorithmisdeterminedbytwomeasures.oneiseciencywhichisreectedbythenumber intorealperformancebenetsinloadbalancingornot.theperformanceofaloadbalancing globalbalancingatruntime.however,fortheotherkindsofapplicationsthatneedtoachieve loadsharingratherthanglobalbalancing,weneedanothermeasure,thebalancequality,to indierentcommunicationmodels. thequestionconcerningtheperformanceofthediusionandthedimension-exchangemethods reecttheabilityofthealgorithminboundingthevarianceofprocessors'workloadsafter performingoneormoreloadbalancingoperations.theobjectiveofthisstudyistoanswer ofattentionfromboththeoreticalandexperimentalresearchers.thediusionmethodwas rstmodeledusinglinearsystemtheorybycybenko[8],andbertsekasandtsitsiklis[7].cybenkoshowedthatthediusionmethodwilleventuallycoerceanyinitialworkloaddistribution Intheliterature,thediusionandthedimension-exchangemethodshavereceivedalot intoaglobaluniformdistributioninstaticsituationsinwhichnoworkloadsaregeneratedor consumedduringloadbalancing,andpresentedanasymptoticboundforthevarianceofany workloaddistributionduringloadbalancinginthedynamicsituation.similarconvergence resultsinthestaticsituationwereobtainedindependentlybyboillat[2].boillatalsoproved ThediusionmethodinthedynamicsituationwasstudiedbyHongetal.[3],andQianand Yang[4],aswell.Theypresentedaconstantboundforthevarianceofworkloaddistribution thatthediusionloadbalancingwillconvergetoaglobalbalancedstateinpolynomialtime. itsoptimalvaluesforthemeshandthetorusnetworks[5]. Lauanalyzedtheeectsoftheparameterontheeciencyofthediusionmethod,andderived whenapplyingthemethodtosomespecicstructures.thediusionmethodischaracterized byaparameterwhichdeterminestheportionofexcessworkloadtobediusedaway.xuand benkoshowedthatregardlessoftheorderofdimensionsconsidered,thissimpleloadbalancing allelcomputers,inwhichbalancingproceedsiterativelyindimensions.ateachdimension, aprocessorbalancesitsworkloadwiththatofitsneighborbelongingtothedimension.cy- Thedimension-exchangemethodwasconceptuallydesignedforhypercube-structuredpar- methodyieldsauniformdistributionfromanyinitialworkloaddistributionafteraroundof balancingoperations[8].healsorevealedthesuperiorityofthedimension-exchangemethod 3

4 overthediusionmethodintermsoftheirecienciesandbalancequalities. appliedthismethodtoarbitrarystructuresbasedonedge-coloring[6].furthermore,xuand Laushowedthat\equalsplitting"oftheworkloadinapairwisebalancingoperationmight notleadtomaximumeciencyinmostpopularstructures,suchasthemeshandthetorus, Thedimension-exchangemethodisnotlimitedtohypercubestructures.Hosseinietal. formforthen-dmeshandtorusstructures. althoughitperformsbestinthehypercube[5,9].throughintroducinganexchangeparameter togovernthesplittingofworkloadateverystep,theyderivedtheoptimalvaluesinclosed theirsoundmathematicalfoundation.onthepracticalside,thebenetsofthediusion methodweredemonstratedinthecontextofdistributedcomputationsofbranch-and-bound algorithms[7,4],andthedimension-exchangemethodwasexperimentedinparallelgraph Thetheoreticalstudyofthediusionandthedimension-exchangemethodsestablished partitioning[8]andperiodicre-mappingofdataparallelcomputations[9].also,willebeek- concludedthatthespeedupduetothedimension-exchangemethodisbetterthanthespeedup LeMairandReeves[4]comparedtheresultsofthesetwomethodsinthedistributedcomputationofbranch-and-boundalgorithmsonahypercube-structurediPSC/2.Theirexperiments ofthedimension-exchangemethodinhypercubes,itmightnotbethecaseforotherpopular duetothediusionmethod.itisinagreementwiththecybenko'sresult. networks.ontheotherhand,previoustheoreticalstudiesofthesetwomethodsweremostly ontheirsynchronousimplementationsinwhichallprocessorsparticipateinloadbalancing Althoughtheresultsofboththeoreticalandexperimentalstudypointtothesuperiority resultshavebeenobtainedontheasynchronousimplementationsofthesemethods.bertsekas workloadmigrationsdemandedbythecurrentoperationhavecompleted.relativelylittle andtsitsiklisprovedtheconvergencepropertyofanasynchronousimplementationofthe operationssimultaneouslyandeachprocessorcannotproceedintothenextstepuntilthe diusionmethod[7],andsongextendedtheresulttothecaseofthetotalworkloadbeingtoo smalltobedividedinnitely[20].lulingandmonienconsideredarandomizedversionofthe diusionmethodinwhichaprocessorinneedofloadbalancingactivatesanoperationamong boththeissuesofeciencyandbalancequalitytogether. dierencebetweenanytwoprocessorsbounded[2].however,noneoftheseworksaddressed anumberofrandomlychosenneighbors,andshowedthatthealgorithmwillkeeptheworkload cationpolicies,andwithstatic/dynamicrandomworkloadbehaviors.thecommunication exchangemethodsintermsoftheireciencyandbalancingqualitywhentheyareimplemented inbothone-portandall-portcommunicationmodels,usingsynchronous/asynchronousinvo- Inthispaper,wemakeacomprehensivecomparisonbetweenthediusionandthedimension- cases:thering,thechain,thehypercubeandthek-aryn-cube.themeshandthetorusallow networkstobeconsideredincludethestructuresofn-dtorusandmesh,andtheirspecial dierentnumberofnodesindierentdimensions.ak-aryn-cubeisaspecialcaseofthen-d torusinthatithasknodesineachdimension[22,23].thehypercubeisaspecialcaseofboth then-dmeshandthek-aryn-cube.ahypercubeisann-dmeshhavingtwonodesineach 4

5 themostpopularchoicesoftopologiesincommercialparallelcomputers[23,24]. dimension,thatis,a2-aryn-cube.welimitourscopetothesestructuresbecausetheyare oftheparametervalueineachmethod:theaveragedimension-exchange(ade),theoptimallytuneddimension-exchange(ode),thelocalaveragediusion(adf),andtheoptimally-tuned Boththedimension-exchangeandthediusionmethodsareparameterizedmethods.Their performanceislargelyinuencedbythechoiceoftheparametervalues.wefocusontwochoices diusion(odf).theoptimalityhereisintermsoftheeciencyinstaticsynchronousimplementationsamongvariouschoicesofthedimension-exchangeandthediusionparameters. Theaverageversionsarethemostoriginalversionswhenthemethodswererstproposedand arestillbeingemployedinrealapplicationstoday.ourmainresultsarethatthedimensionexchangemethodoutperformsthediusionmethodintheone-portcommunicationmodel;in balancingevenundertheall-portcommunicationmodel;thestrengthofthediusionmethodis inasynchronousimplementationundertheall-portcommunicationmodel;theodfalgorithm particular,theodealgorithmisfoundtobebestsuitedforsynchronousimplementationin performsbestinthiscase. thestaticsituation;andthatthedimension-exchangemethodissuperiorinsynchronousload Section3describestheloadbalancingalgorithmsinauniedform.InbothSection4andSection5,thealgorithmsarecomparedwithrespecttotheirimplementationusingasynchronous insection2,whichprovidesaframeworkforthecomparisonoftheloadbalancingalgorithms. Therestofpaperisorganizedasfollows.Werstpresentagenericmodelofloadbalancing ancingalgorithms.weconcludeinsection7withasummaryoftheresultsofthecomparison betweenthedimension-exchangeandthediusionmethods. whichverifyourtheoreticalresultsaswellasprovidefurtherinformationontheseloadbal- andsynchronousinvocationpolicies,respectively.section6givestheresultsfromsimulations, cessorsinterconnectedbyadirectcommunicationnetwork.processorscommunicatethrough 2Weconsideraclassofparallelcomputerswhicharecomposedofanitesetofhomogeneouspro- Agenericmodelofloadbalancing passingmessages.thecommunicationchannelsareassumedtobefullduplexsothatapair ofdirectlyconnected(nearestneighbor)processorscansend/receivemessagessimultaneously andevvisasetofedges.everyedge(i;j)2ecorrespondstothecommunication messagesthroughachannelcantakeplaceinstantaneously.werepresentsuchasystemby asimpleconnectedgraphg=(v;e),wherevisasetofprocessorslabeledthroughn, to/fromeachother.inaddition,weassumethattheoperationsofsendingandreceiving channelbetweenprocessorsiandj.leta(i)denotethesetofnearestneighborsofprocessor i,d(i)=ja(i)jbethedegreeofprocessori,anddbethemaximumofd(i)forin. tobelargeenoughsothattheworkloadofaprocessorisinnitelydivisible.processesmaybe processes,whicharethebasicunitsofworkload.thetotalnumberofprocessesareassumed Theunderlyingparallelcomputationisassumedtocomprisealargenumberofindependent 5

6 dynamicallygenerated,consumed,ormigratedduetoimbalanceasthecomputationproceeds. ispossiblewhenprocessorsarecapableofmultiprogrammingorthebalancingoperationisdone operation,orbothoperationssimultaneously.theconcurrentexecutionofthesetwooperations operation.ananytime,aprocessorcanperformacomputationaloperation,abalancing Weclassifytheoperationsintotwotypes:thecomputationaloperationandthebalancing inthebackgroundbyspecialcoprocessors.theworkloadofprocessorscanbeeitherxedor varyingwithtimeduringtheloadbalancingoperation,whichwerefertoasthestaticandthe processoriattimetbywtiintermsofthenumberofresidingprocesses.weuseintegertime dynamicsituations,respectively. tosimplifythepresentation.theresultscanbeexpendedreadilytocontinuoustime.leti(t) denotethesetofprocessorsperformingbalancingoperationsattimet.thechangeofworkload Lettbeatimevariable,representingglobalrealtime.Wequantifytheworkloadof ofaprocessorattimetcanbemodeledbythefollowingequationinthestaticsituation andthefollowingequationinthedynamicsituation wt+ i=(wti+t+ fi(wti;wtjjj2a(i))i2i(t) i i62i(t) () wheret+ wt+ i=(wti+t+ i denotestheamountofworkloadgeneratedornishedfromtimettot+,and fi(wti;wtjjj2a(i))+t+ i i i2i(t) i62i(t) (2) fi()representsaloadbalancingoperator. loadbalancingatanytimet,i(t),areleftunspecied.theoperatorfi()canbeanynearestneighborloadbalancingalgorithm,includingthediusionandthedimension-exchange methods.theseti(t)isdeterminedbytheinvocationpolicyoftheloadbalancing.the Thismodelisgenericbecausetheloadbalanceoperatorfi()andthesetofprocessorsin choiceofi(t)isorthogonaltotheloadbalancingalgorithminthatanyinvocationpolicy canbeusedincombinationwithanyloadbalancingalgorithminimplementation.sincea parallelcomputationsusingdomaindecompositiontechniques,forexample,thecomputational requirementassociatedwitheachportionofaproblemdomainmaychangeasthecomputation loadbalancingoperationincursnon-negligibleoverheads,dierentapplicationsrequiredierentinvocationpoliciesforabettertradeobetweenperformancebenetsandoverheads.in proceeds.aneectivewaytoreducethepenaltyduetoloadimbalancesistoperiodicallyredecomposetheproblemdomainwiththeaimofachievingaglobaluniformdistributionacross theprocessors.tothisend,allprocessorsarerequiredtoperformloadbalancingoperations Bycontrast,theparallelexecutionofdynamictree-structuredcomputationsusuallyrequires theinstancewhentheglobalsystemstatesatisescertainconditionssuchasthosesetin[25]. synchronouslyforashorttimeperiod.thatis,i(t)=f;2;;:::;ngfortt0,wheret0is sors.thus,eachprocessorisallowedtoinvokealoadbalancingoperationatanytimeinan onlyloadsharing assuringthatnoidleprocessorsexistwhilethereareotherbusyproces- asynchronousmanneraccordingtoitsownlocalworkloaddistribution.asimplepolicyisto 6

7 5 5 Processor 4 4 Processor activatealoadbalancingoperationonceaprocessor'sworkloaddropsbelowapresetthreshold,wunderload,i.e.,i(t)=fijwti<wunderloadg.moresophisticatedinvocationpolicieswere discussedin[2,4].inshort,wemakeadistinctionbetweensynchronousandasynchronous implementationsofloadbalancingaccordingtotheirinvocationpolicies.figurepresents respectively. oneexampleofthesetwoimplementationmodelsinasystemofveprocessors.thedots andthetrianglesrepresentthecomputationaloperationsandtheloadbalancingoperations, t t+5 t+0 t+5 time t t+5 t+0 t+5 time cedure.weareconcernedwithsubsequentworkloaddistributionsresultingfromdierent Figure:Anillustrationofgenericmodelsofloadbalancing loadbalancingalgorithms.denotetheoverallworkloaddistributionatcertaintimetbya Assumet=0whenprocessorsinvokeasynchronousorasynchronousloadbalancingpro- (a) Asynchronous implementation (b) Synchronous implementation Wt=(wt;wt;;wt),wherewt=PNi=wti=N.Wedenetheworkloadvariance,denotedby t,asthedeviationofwtfromwt;thatis, vectorwt=(wt;wt2;;wtn).denoteitscorrespondinguniformdistributionbyavector Withtheworkloadvariancet,wedenetheeciencyofaloadbalancingalgorithmasthe t=jjwt?wtjj2=nxi=(wti?wt)2: numberofloadbalancingstepsrequiredtoreducethevarianceoftheinitialstatetoatolerable Throughoutthepaper,E[]denotestheexpectedvalueofarandomvariable. algorithmswillbecomparedintermsofthesetwomeasuresunderthefollowingassumption. istobeguaranteedbytheloadbalancingprocedureinthedynamicsituation.loadbalancing levelinthestaticsituation;anddenethebalancequalityastheboundforthevariancewhich timet,t0,processors'workloadgeneration/consumptionamount,ti,in,arezero Assumption2.Initially,processors'workloads,w0i,iN,areNindependentand identicallydistributed(i.i.d.)randomvariableswithexpectation0andvariance20.atany inthestaticsituationori.i.d.randomvariableswithexpectationandvariance2inthe dynamicsituation. 7

8 Thissectionbrieydescribesthedimension-exchangeandthediusionmethods.Bothof themareparameterizedloadbalancingalgorithms.wepresentseveralinstancesofthesetwo 3 Thedimension-exchangeandthediusionmethods methodsbasedondierentchoicesofvaluesfortheirparameters. 3.Thedimension-exchangemethod way Withthedimension-exchangemethod,anyprocessorwhichinvokesaloadbalancingoperation balancesitsworkloadwithitsneighborssuccessively.foraprocessori,itworksinthefollowing f()for(c=;cd(i);c++) valuebeforehandwhichdeterminesthefractionofexcessworkloadtobemigratedbetweena wherejc2a(i);and0<<,calledthedimension-exchangeparameter,isgivenaxed wi=wi+(wjc?wi) (3) balancing.itisbecauseofthesequentialnatureinthesequenceofbalancingsteps,aload balancesitsworkloadwithoneofitsneighbors,andusesthenewresultforthesubsequent methodcomprisesd(i)pairwisebalancingstepsforprocessori.ateachstep,processori pairofprocessors.theformulasaysthatabalancingoperationinthedimension-exchange balancingoperationrequiresd(i)communicationstepsinboththeall-portandtheone-port communicationmodels. choicesoftheparameterwhichhavebeensuggestedasrationalchoicesintheliterature. parameter.adimension-exchangeoperationwithdierentchoicesoftheparameterwillreduce theworkloadvarianceofthesystembydierentdegrees.inthefollowing,wepresenttwo Theeciencyofthedimension-exchangemethodisdeterminedbythedimension-exchange.Averagedimension-exchange(ADE)equallysplitsthetotalworkloadofapairofprocessors 2.Optimally-tuneddimension-exchange(ODE)takescertainspecicparametervaluesthat operation,andhasbeenfavoredinhypercube-structuredsystems[8,26,27]. thatis,==2.itisastraightforwardchoiceforlocalbalancingateachpairwise havetheeectofmaximizingeciencyinstaticandsynchronousbalancing[5,9].the optimalparameterdependsonthetopologyandthesizeofunderlyingcommunication network.letk=maxfki;inginthekk2knmeshandtorus.then,their optimalparametervalueswereshown,in[9],tobe ==(+sin(2=k))inthetorus. ==(+sin(=k))inthemesh, onlyafewprocessorsthatarenotclosetoeachotherareinneedofloadbalancingatthe sametime.however,itssynchronousimplementationrequiresprocessorstobecoordinatedin Thedimension-exchangemethodcanbeimplementedwithoutdicultyincaseswhere 8

9 avoidcommunicationcollisions.thepotentialofparalleleciencyisduetothefactthatthe ordertoparallelizebalancingoperationsalongdierentcommunicationchannelsaswellasto executionorderofpairwisebalancingstepsintheoperationf()ofeq.(2)isleftundened. Thepairwisebalancingstepsalongthechannelsinthesamesubsetcanthenbeperformed ofedgesintoanumberofsubsetssuchthatnotwoadjoiningedgesareinthesamesubset. Theparallelizationofpairwisebalancingoperationscanberealizedbypartitioningtheset concurrentlywithoutcollisions.suchgraphpartitionisequivalenttotheproblemofedge coloringofgraphs[28].figure2showsexamplesofcolorgraphsofameshandatorus. Thenumbersinparenthesesaretheassignedchromaticindices.Analternativeapproachto parallelizingloadbalancingoperationsisrandommatchingwhichwasusedin[29] Thediusionmethod Figure2:Examplesofcoloredmeshandtorus (a) Colored Mesh (b) Colored torus Withthediusionmethod,anyprocessorwhichinvokesaloadbalancingoperationcompares itsworkloadwiththoseofitsnearestneighbors,andthengivesawayortakesincertainamount canbewrittenintheform ofworkloadwithrespecttoeachofnearestneighbors.thediusionoperatorinaprocessori where0<ij<,calledthediusionparameter,ispredenedtodictatetheportiontobe fi()wi+x balancingoperationwiththediusionmethodrequiresonlyonecommunicationstepinthe migratedbetweenanytwoprocessors.processoriapportionsexcessworkloadjwi?wjjto all-portcommunicationmodel,butd(i)stepsintheone-portcommunicationmodel. processorjifwi>wj,orfetchessomeworkloadfromprocessorjotherwise.clearly,aload bythediusionparameter.followingaretwocommonchoicesoftheparameter..localaveragediusion(adf)takesanaverageoftheworkloadofneighboringprocessors Asinthedimension-exchangemethod,theeciencyofthediusionmethodisdetermined j2a(i)ij(wj?wi) (4)

10 bysettingij= weuseasinglevalue= samedegree.themeshisapproximatelyregularwhenitssizeislarge.forsimplicity, +d(i)[2,3,4];thetorusisregularinthatallprocessorshavethe 2.Optimally-tuneddiusion(ODF)takescertainspecicparametervaluesformaximizing thetorus. +dtocoverallcommunicationchannelsinthemeshandin eciencyinstaticandsynchronousbalancing[8].asinthedimension-exchangemethod, network.letk=maxfk;k2;;knginthekk2knmeshandtorus.then,their theoptimaldiusionparameterdependsonthetopologyandthesizeoftheunderlying optimalchoiceswereshown,in[8,5],tobe ==(n+)inthen-dhypercube. ==(2n+?cos(2=k))inthetorus, ==2ninthemesh, Inanasynchronousimplementationofloadbalancing,processorsperformbalancingoperations 4discretelybasedontheirownlocalworkloaddistributionsandinvocationpolicies.Sinceload Asynchronousimplementations balancingalgorithmscanbetreatedasorthogonaltoinvocationpolicies,weconsidertheload balancingoperationsoftheprocessorsinonetimestepsoastoisolatetheireectsonthe processorisperformingloadbalancingoperations.thedynamicsituationpresentsonlyafew loadbalancinginwhichtheunderlyingcomputationinaprocessorissuspendedwhilethe workloadvariancefromtheeectsofinvocationpolicies.wefocusonthestaticsituationof relativelyminordierencestotheanalysisoftheeectsofloadbalancing. aretheresultsfromvariousloadbalancingoperations. variancewhent=.ourcomparisonwillbemadebetweenade,ode,adf,andodfwhich Let0betheoriginalsystemworkloadvariancewhent=0,andbethesystemworkload Assumption2..Then,E[ade]E[df]intheone-portcommunicationmodel,whileE[df] Theorem4.Supposeprocessorsarerunninganasynchronousloadbalancingprocessunder E[ade]intheall-portcommunicationmodel.Moreover,E[adf]E[odf]inchainandring networks,bute[odf]e[adf]intwo-orhigher-dimensionalmeshesandtori.inaddition, E[ade]E[ode]intheall-portcommunicationmodel. theone-portandtheall-portcommunicationmodels,respectively.morespecically,itreveals thattheodfalgorithmoutperformstheadfalgorithminhigherdimensionalmeshesand torialthoughtheodfwasoriginallyproposedforuseinsynchronousglobalbalancing. Thistheoremsaysthatthedimension-exchangeandthediusionmethodsaresuitablefor ThecalculationofE[]isbasedonalemmaconcerningthesamplevarianceofacombination ThetheoremisprovedthroughthederivationoftheclosedformofeachvarianceE[]. 0

11 ofrandomvariablesinasampleset,whichwepresentwithoutproof.itcanbeeasilyshown =PNi=i.Then, usingfundamentalstatisticaltheories. Lemma4.Supposethat;2;:::;NareNi.i.d.randomvariableswithvariance2,and.foranyk,kN, where0<ai<satisespki=ai=;andthevarianceisminimizedatai==kfora givenk. E(jkXi=aii?j2)=(kX=a2i?N)2; (5) 2.foranykandk2andkk2N, where0<ai<satisespk E(jk i=ai=and0<bj<satisespk2 Xi=aii?j2)E(jk2 Xj=bjj?j2) j=bj=. (6) ProofofTheorem4.Atcertaintimeinanasynchronousloadbalancingprocess,there simultaneously.letea(i)=fig[a(i)denotethebalancingdomainofaninvokerprocessor mightbemorethanoneprocessorthatareinvokingloadbalancingwithintheirneighborhoods areunionsofoverlappingdomains.processorsindierentspheresperformloadbalancing other.asawhole,thoseprocessorsthatarerunningloadbalancingprocessesarepartitioned i.thebalancingdomainsofconcurrentinvokersmayoverlapormaybeseparatedfromeach intoanumberofseparatedspheres,someofwhicharesingularbalancingdomainsandsome operationsindependently,whileprocessorsinthesamesphereperformloadbalancingina B;B2;;Bm.Then,bythedenitionoftheworkloadvarianceandAssumption2.,we synchronousmanner. havesupposeinitiallytherearemindependentbalancingspheresinthesystem,denotedby E[]=E(NXi=jwi?wj2) =NXi=E(jwi?wj2) =mxj=x i2bie(jwi?wj2)+x i2bje(jwi?wj2)+(n?n0)(?n)(20+2); i62[mj=bje(jwi?wj2) wheren0=j[mi=bijisthenumberofprocessorsinvolvedinloadbalancing.thelastterm (7) of(7)isduetotheunderlyingcomputationaloperations.itisaconstantforagivenn0and

12 independentofthetopologicalrelationshipsamongthen0processors.thersttermof(7)is thattheexpectedvalueofthesystemworkloadvarianceisinuencedindependentlybyload duetoloadbalancingoperationsinallseparatedbalancingspheres.itisasimplearithmetic sumofworkloadvarianceofeachsphere,pi2bje(jwi?wj2).asawhole,eq.(7)implies balancingoperationswithindierentbalancingspheres.therefore,itsucestocomparethe Case:loadbalancinginasingularbalancingdomain eectsofloadbalancingalgorithmswithindierentspheresusinglemma4.. Werstconsiderloadbalancinginspheresofsingularbalancingdomains.SupposeBissuch B.Then,withthediusionalgorithm,theworkloadsofprocessorsattheendofadiusion asphere,andwithoutlossofgenerality,b=ea()=f;2;3;;d+g.thatis,processor operationaregiven,accordingtoeq.(4),by invokesaloadbalancingoperationwithinitsdneighborswhicharelabeledfrom2tod+. LetX=Pd+ i=e(jwi?wj2),denotingtheexpectedvalueofworkloadvarianceofsphere InvokingLemma4.oneachcomponentwi,wehavethat wi=((?d)w0+pd+ w0+(?)w0ij=2w0iifi=; if2id+; (8) Xdf=d+ =E(j(?d)w0+d+ Xi=E(jwi?wj2) =[d2+(?d)2?=n]20+d[2+(?)2?=n)]20 Xi=2w0i?w0j2)+d+ Xi=2E(jw0+(?)w0i?w0j2) =(d22+3d2?4d+d+?d+ Letopt=2=(3+d).WereplacebyoptintheexpressionofXdf,andobtain ItcanbeeasilyveriedthatXdf,asaconvexfunctionof,isminimizedat=2=(3+d). N)20: (9) Recallthatadf==(d+),andthatodf==dinameshandodf==(d+?cos(2=k)) Xdfj(opt)=(d2+3 d+3?d+ inatorus,wherekisthemaximumdimensionalorderofthetorus.itfollowsthat N)20: (0) inthecaseofachain(i.e.,the-dmesh)whered=2,adf<2=(3+d)<odfand inthecaseofaring(i.e.,the-dtorus)whered=2,adf<opt<odfandjadf? jadf?optj<jodf?optj; inthecaseofhigherdimensionalmeshesandtoriwhered4,adf<odf<opt. optj<jodf?optj,fork2; 2

13 Consequently,withthediusionmethod, (XodfXadfXdf(opt)ifd=2andk2; Withthedimension-exchangemethod,processorisassumedtoperformpairwiseloadbalancingwithprocessors2;3;:::;d+inturninadimension-exchangeloadbalancingoperation. XadfXodfXdf(opt)ifd4 () theendofadimension-exchangeoperationaregiven,accordingtoeq.(3),by Assumetheunderlyingsystemisintheone-portcommunicationmodel.Then,theworkload generation/consumptionratioinaroundofpairwisebalancingstepshasthesamestatistical characteristicsasthoseinadiusionoperation.consequently,theprocessors'workloadsat wi=8><>:(?)dw0+pd? (?)w0i+(?)i?2w0+2pi?3 (?)w02+w0j=0(?)jw0d?j+ j=0(?)jw0i??jif3id+; ifi=2; ifi=; InvokingLemma4.oneachcomponentwi,wehavethat (2) Xde=d+ =[(?)2d+2d? Xi=E(jwi?wj2) +d+ Xi=3[(?)2+2(?)2(i?2)+4i?3 Xj=0(?)2j?=N]2o+[(?)2+2?=N]20 =[d(?)2+22?(?)2d?(?)2+4d??(?)2?(?)2d?2 Xj=0(?)2j?(d?)=N]20 Inparticular,substituting=2fortheintheexpressionofXde,callingitXade,leadstothat +(?)2d?d+ N]20: Xade=(3d+5+22?2d 9?d+ communicationmodel, From(0)and(3),itisknownthatXadeXdf(opt):Itisthusprovedthatintheone-port N)20: (3) asmuchtimeasadiusionloadbalancingoperation.thatis,inatimestepofthediusion Intheall-portcommunicationmodel,adimension-exchangepairwisebalancingsteptakes XadeXdf: (4) method,aprocessorbalanceswithonlyoneofitsneighborswiththedimension-exchange method.itresultsinthat Consequently,XadeislessthanXodebutlargerthanXdf. Xde=2[(?)2+2?N]20+(d?)(?N)20: 3

14 Case2:loadbalancinginaunionofoverlappingdomains Wenowconsiderloadbalancinginsphereswhichareunionsofoverlappingbalancingdomains. Abalancingspherecanbeaunionofanynumberofoverlappingdomains.Inconsideration ofthelikelihoodthatfewprocessorswillbeinvokingloadbalancingsimultaneouslyinasynchronousimplementations,wefocusontheunionoftwobalancingdomainsonly.figurepingbalancingdomainsin2-dmeshesandtori.thetrianglesareinvokersofloadbalancing illustratesthreetopologicalrelationshipsbetweenapairofprocessorswhichhaveoverlap- processesandthedotsareprocessorsbeinginvolvedinloadbalancing. Figure3:Illustrationsofoverlappingbalancingdomains wj2).supposeprocessorsjandj2havethesamenumberofdirectneighbors.then,inthe ea(j2).letydenotetheexpectedvalueoftheworkloadvarianceofb2,i.e.,y=pi2b2e(jwi? AssumeB2isaunionofbalancingdomainsofprocessorsjandj2.Thatis,B2=eA(j)[ (a) (b) (c) casethatprocessorsjandj2aredirectlyconnected,asinfigure3(a),wehavethatinthe andinthedimension-exchangemethod, Inparticular, Yde2Xde?2[(?)2+2(?)2(d?)+4?(?)2(d?) Yade2Xade?2(3+ 2?(?)2?N]20: (6) (orea(j2)nfjg)changesitsworkloadinthesamewayasinloadbalancingwithinasingular balancingdomainea(j)(orea(j2)).thereasonsoftheinequalityoftheydeinthedimensionexchangemethodareasfollows.withthedimension-exchangemethod,bothprocessorsjand TheequationofthediusionmethodisduetothefactthateachprocessorineA(j)nfj2g j2performpairwisebalancingoperationswiththeirneighborsinturnaccordingtoorderswhich as,processorj2asc+,andotherneighboringprocessorsofprocessorjas2tod+ inb2isthusinuencedbytheexecutionorderacrossthecommunicationchannels.suppose thechannel(j;j2)isindexedascth.withoutlossofgenerality,werelabeltheprocessorj arepresetthroughedge-coloringofthesystemgraph.thechangeoftheworkloaddistribution 4 diusionmethod, Ydf=2Xdf?2[(?)2+2?N]20; (5) 322d?N)20: (7)

15 excludingc+.then,itisclearthatprocessorsfrom2tocchangetheirworkloadsin thesamewayastheircounterpartsiftheyareperformingloadbalancingwithinasingular E(jwd+?wj2)E(jwi?wj2)fori>c,theboundofYdeishenceobtained. domainea(i)alone,whilethebehaviorsofotherprocessorswillbeinuencedbyprocessorsin workloadvariancee(jwi?wj2)inaunionofoverlappingdomainsthaninea(i)alone.since ea(j)nfig.fromlemma4.(2),itisalsoknownthateachprocessori,i>c,willpossessless theoptimalchoiceof.then,ydf(opt)=d?4d2?4d+ ItcanbeeasilyveriedthatYdfisminimizedat=(2d?)=(d2+3d?2).Letoptbe Asinthecaseofsingularbalancingdomain,itcanbeshownthat d2+3d?2: (8) inthecaseofachain(i.e.,-dmesh)whered=2,adf<opt<odfandjadf?optj< inthecaseofaring(i.e.,-dtorus)whered=2,adf<opt<odfandjadf?optj< jodf?optj; inthecaseofhigherdimensionalmeshesandtoriwhered4,adf<odf<opt. jodf?optj,fork6; Consequently,withthediusionmethod, (YodfYadfYdf(opt)ifd=2andk6; YadeYdf(opt). Ontheotherhand,thecomparisonbetweenYadeofEq.(7)andYdfj=optofEq.(8)yields YadfYodfYdf(opt)ifd4 (9) thereareatmosttwoprocessorsintheintersectoftheirbalancingdomainsinthemeshand torusnetworks.letsbethecardinalityoftheintersectea(i)\ea(j),s=or2.then,with thediusionmethod, Incasesthatprocessorsiandjarenonadjacent,asillustratedinFigure3(b)and3(c), Ydf=2Xdf?2s[(?)2+2?N]20+s[(?2)2+22?N]20 andwiththeadealgorithm, =2Xdf?s[(?22)?N]20; (20) Yade2Xade?s[3+ 322d?N]20: 2 YodfYadfincased=2,andYadfYodfincased4. Similarlytothecaseofsingularbalancingdomain,wehavetheresultthatYadeYdf, (2) ontheassumptionofone-portcommunicationmodel.intheall-portcommunicationmodel, Noticethattheprecedinganalysisofthedimension-exchangemethodisimplicitlybased 5

16 adimension-exchangepairwisebalancingstepcorrespondstoadiusionbalancingoperation. formedconcurrently,wethushave Becausetwopairwisebalancingoperationsinaunionoftwobalancingdomainscanbeper- wheres=or2.obviously,yadeislessthanyodebutlargerthanydf. Thetheoremisthenproved. Yde=4[(?)2+2?N]20+(2d?4?s)(?N)20; theoremstillholdsinthedynamicsituation.considerprocessorsinbalancingsphereb.since theworkloadsgenerated/consumedfromtime0totimeinanyprocessori,i2b,willnotbe consideredinitsloadbalancingoperationattimestep,theoperationinthedynamicsituation Notethateventhoughtheproofofthetheoremassumesstaticworkloadbehaviors,the N0=jBj.TheaddedtermisaconstantforagivenN0andindependentoftheloadbalancing workloadvarianceinthestaticsituation.asawhole,theaccumulativeworkloadvarianceof processorsinbalancingspherebinthedynamicsituationispi2be(wi?wj2)+n02,where thenresultsinanworkloadvariancee(jwi?wj2)+2,wheree(jwi?wj2)istheprocessor's situation. algorithmused.hence,theargumentsintheproofofthetheoremarevalidinthedynamic workloadofaprocessor,sayprocessor,anditssurroundingdprocessors,labeledfrom2to d+,inasimplewaythatprocessorgives(w?wi)loadstoprocessori,inthecaseof w>wi,andtakes(wi?w)loadfromprocessori,otherwise(2id+).inasingular Toconcludethissection,weremarkthatadiusionloadbalancingoperationaveragesthe balancing.specically,processorcalculatesthelocalaveragewas balancingdomain,theremightbeavariantoftheadfalgorithmwhichstrivesforlocalload w=w+p2id+wi iisdecientornot.aftersuchanoperation,eachprocessori,2id+endsupwiththe andthengivesortakesjwi?wjloadstoorfromprocessoriaccordingtowhetherprocessor +d ; Pd+ sameworkloadasprocessor.consequently,theexpectedworkloadvarianceofthedomain i=e(jwi?wj2)becomes model.althoughitincursmoreoverheadsthananodforadfoperation,suchavariantof whichisobviouslysmallerthanthatoftheademethodevenintheone-portcommunication (?d+ N)20; inbalancingsphereswhereanumberofbalancingdomainsoverlapwitheachotherbecause processorsinsuchasphereareunabletobalancetheirworkloadswithalltheprocessorsin suchanoperation. diusionoperationispreferredinsingularbalancingdomains.however,itmaynotbeeective 6

17 5Inasynchronousimplementationofloadbalancing,allprocessorsperformloadbalancing operationsconcurrentlyandcontinuouslyforatimeperiodinordertoachieveaglobalbalanced Synchronousimplementations dynamicsituation. stateinthestaticsituationortokeepthevaryingsystemworkloadvarianceboundedinthe oftheworkloaddistributionattimetinthediusionmethodcanbemodeledbytheequation bemodeledbylineariterativeprocesses,asillustratedin[2,8,5].fromeq.(4),thechange Thesynchronousimplementationofthediusionandthedimension-exchangemethodscan whered,calledadiusionmatrix,isgivenby Dij=8><>: ifprocessorsiandjaredirectlyconnected; Wt+=DWt+t; (22) 0?d(i)ifi=j; methodarefullycapturedbytheiterativeprocessgovernedbyd. Withtheaboveformulation,thefeaturesofthesynchronousimplementationofthediusion otherwise: methodcanbemodeledbytheequation Then,fromEq.(3),thechangeoftheworkloaddistributionattimetinthedimension-exchange Letbetheminimumnumberofcolorsrequiredforedgecoloringofthesystemgraph. wherem,calledthedimension-exchangematrix,isgivenby Wt+=MWt+t M=MM?:::M: (23) EachMc(c)reectsthechangeoftheworkloaddistributionofthesystematpairwise balancingstepcoftimet.thus,thefeaturesofthesynchronousimplementationofthe dimension-exchangemethodarefullycapturedbytheiterativeprocessgovernedbym. balancingoperationssimultaneouslyandallcomputationaloperationsaresuspended.this 5.Staticsituation situationistrueofperiodicloadbalancing,asexperimentedin[30,3,25,9].theeciency ofaloadbalancingalgorithminthissituationisreectedbythenumberofcommunication Inastaticsynchronousloadbalancingprocess,allprocessorsareassumedtoperformload stepsrequiredforarrivingataglobalbalancedstatefromanyinitialloaddistribution. FtWt=Wt,itfollowsthat Wt=FtW0,whereFt=FFF LetFbeeitherthedimension-exchangematrixMorthediusionmatrixD.Then, ttimes {z }.SinceWt=W0inthestaticsituation,and Wt?Wt=F(Wt??Wt?)=Ft(W0?W0): 7

18 Then,bythedenitionoftheworkloadvariance,wehave where(f)isthesub-dominanteigenvalueoffinmodulus.itsaysthattheworkloadvariance isreducedgeometrically,anditsscalefactorisupperboundedby(f).theboundistight, t=jjwt?wjj2=jjft(w0?w)jj22t(f)0; Thesub-dominanteigenvalueinmodulus(F)isthusreferredalsoastheconvergencefactor andtsatises ofaloadbalancingalgorithm. t'2t(f)0 forlarget. (24) initialstatetosomeprescribedbound.then,fromeq.24,itfollowsthat LetTbethenumberofiterationstepsrequiredtoreducetheworkloadvarianceofan Hence, T=ln?ln0 T=O(=ln(F)): 2ln(F): (26) (25) algorithm,adf,whenappliedtoabroadvarietyofstructures.in[5,5],xuandlauanalyzed atedbyanumberofresearchers.in[2],boillatpresentedtheconvergencefactorsoftheadf theeectsofthedimension-exchangeandthediusionparametersontheeciencyofload Theconvergencefactorsofthedimension-exchangeandthediusionmethodswereevalu- balancing,andproposedtheodeandodfalgorithmsbychoosingtheoptimalvaluesforthe parametersand.thecorrespondingconvergencefactors,odeandodf,arehencereadily work.wesummarizetheconvergencefactorsintable. availablefromtheirproofs.also,theconvergencefactoradecanbederivedeasilyfromthe isthemaximumnumberofnodesoveralldimensionsofann-dnetwork Table:Convergencefactorsofthedimension-exchangeandthediusionmethods,wherek toruscos2(2=k)?sin(2=k) ADE DEmethod ODE ADF Diusionmethod meshcos2(=k) +sin(2=k)2n?+2cos(2=k)?sin(=k) +sin(=k) 2n?+2cos(=k) 2n+ 2n?+cos(2=k) 2n+?cos(2=k) n?+cos(=k) ODF Noticethattheconvergencefactorisiniterationsteps,eachofwhichiswhatwecalleda n requiresonlyonecommunicationstepwhileadimension-exchangeoperationstillrequires2n boththedimension-exchangeandthediusionmethodsrequires2ncommunicationstepsin ann-dnetwork.intheall-portcommunicationmodel,adiusionloadbalancingoperation loadbalancingoperationbefore.intheone-portcommunicationmodel,suchanoperationin Byg(t)'h(t)forlarget,wemeanthatg(t)=h(t)?!forlarget. 8

19 steps.therefore,tableandtheeq.(26)leadtotable2.itpresentsthetimecomplexities incommunicationstepsnecessaryforvariousloadbalancingalgorithmsinbothone-portand all-portcommunicationmodels. themaximumnumberofnodesoveralldimensionsinann-dnetworkand?portmeansthe all-portcommunicationmodel. Table2:Timecomplexitiesofthedimension-exchangeandthediusionmethods,wherekis toruso(nk2)o(nk2)o(nk)o(nk)o(n2k2)o(nk2)o(n2k2)o(nk2) mesho(nk2)o(nk2)o(nk)o(nk)o(n2k2)o(nk2)o(n2k2)o(nk2) -port ADE*-port-port*-port ODE -port ADF*-port -port ODF*-port example,theo(nk)estimatefortheodealgorithmfollowsfromthefollowingderivation. ThetimecomplexitiesgiveninTable2areinferredfromtheconvergencefactors.For ln(ode)=ln(?sin(2=k) =ln(?2sin(2=k) +sin(2=k)) 'ln(?4 +sin(2=k)) ' k+2 k+2) forlargek FromEq.(26),wehaveTode=O(k)inbalancingoperations.SinceanODEloadbalancing O(nk)isthusproved. operationrequireso(n)communicationstepsinbothcommunicationmodels,theestimate Theorem5.Supposeprocessorsarerunningsynchronousloadbalancingprocessesinthe staticsituation.then,boththeadeandtheodealgorithmsconvergeasymptoticallyfaster Theentriesofthetableshowthefollowing. k. thanthediusionmethodintheone-portcommunicationmodel.intheall-portcommunication model,theodealgorithmconvergesalsofasterthantheotherthreealgorithmsbyafactorof 5.2Dynamicsituation chronousimplementationofthediusionmethodinthedynamicsituationhasbeenevaluated anceofprocessors'workloadsboundedastightlyaspossible.theperformanceofthesyn- computationconcurrently.dynamicloadbalancinginthissituationaimsatkeepingthevari- Indynamicsynchronousimplementations,allprocessorsareperformingloadbalancingand in[8,3,4].in[8],cybenkoshowedthatthediusionmethodkeepstheasymptoticvariance 9

20 this,wearestillunabletodrawaconclusionregardingthesuperiorityoftheadealgorithm thevariancefromtheadealgorithmwhenbothareappliedtothehypercubenetwork.given intermsofthebalancequalityduringloadbalancing.in[3],hong,tanandchenreported bounded.healsoprovedthattheasymptoticvariancefromthediusionmethodislargerthan aconstantboundfortheworkloadvariancewhentheadfalgorithmrunsinthehypercube network.thisresultwasextendedlaterbyqianandyangtogeneralizedhypercubesandmesh structures[4].althoughtheboundstheyderivedareindependentoftime,theyaretooloose tobeusedforthecomparisonofbalancequalitiesduringloadbalancing.also,theapproaches theirdierentoperationalbehaviors. usedin[3,4]areunsuitablefortheanalysisofthedimension-exchangemethodbecauseof algorithms.wepresentaclosedformoftheworkloadvariancewhenaloadbalancingprocess runsinthetorusandthehypercubenetworks.theapproachisnotapplicabletothecase ofthemeshnetworksastheyarenotregularnetworks.nevertheless,sinceann-dmeshhas Inthissubsection,wedevelopanewapproachforanalyzingthebalancequalitiesofdierent onlyafractionofitsprocessorswhosedegreeissmallerthan2n,ourresultsasareasonable resultstobepresentedinthenextsection. approximationcanbeappliedtothemeshstructureaswell;thisissupportedbyoursimulation Lemma5.Supposeprocessorsarerunningasynchronousdiusionloadbalancingprocess d=2nbethedegreeofthenetwork. Throughoutthesubsection,weassumeloadbalancinginann-Dtorusnetwork,andlet underassumption2..then,e(wt)isauniformdistributionatanytimetand wherea=(?d)2+d2. E[tdf]=(at+20+?at+?a2)N?(t+)2?20; (27) Proof.TheuniformdistributionofE(Wt)resultingfromthediusionmethodcanbeeasily shown.weomititsproofbecauseitisalsoavailableasaspecialcaseintheproofoftheuniform distributionofe(wt)resultingfromthedimension-exchangemethodinthenextlemma. haveconsidertheexpectedworkloadvariancee[tdf].byitsdenitionandassumption2.,we E[tdf]=E(jjWt?Wtjj2) =E(jjDt+W0?W0jj2+tXi=0E(jjDit+?i?t+?ijj2) =E(jjDWt??Wt?jj2)+E(jjt?tjj2) =(at+20+?at+ =N(at+?N)20+tXi=0(ai?N)2?a2)N?(t+)2?20; 20

21 nent'sd+sub-componentswithcoecients?d;;;:::;;andasequenceofoperation distributionchangeseachofitscomponentstobecomealinearcombinationofthecompo- wherethefourthstepisbasedonthefollowingobservations.anoperationdontheworkload Dtchangeseachcomponentoftobecomealinearcombinationofits(d+)tsub-components. N(a?=N)2,andE(jjDt?jj2)=N(at?=N)2,wherea=(?d)2+d2. terminedonlybytheircombinatorialcoecients.therefore,wehavee(jjd?t+?ijj2)= FromLemma4.,itisknownthatthevarianceofacombinationofrandomvariablesisde- allpossiblechoicesoftheparameter,whichhappenstobethechoiceoftheadfalgorithm inn-dmeshesandtori.immediately,weobtain Considertheterma=(?d)2+d2inLemma5..Itisminimizedat==(d+)over presentacompaniontolemma5.inthefollowing. Next,weconsidersynchronousimplementationsofthedimension-exchangemethod.We E[tadf]E[todf]: (28) Lemma5.2Supposeprocessorsarerunningasynchronousdimension-exchangeloadbalancingprocessunderAssumption2.,exceptthatprocessorsgenerate/consumeiworkloadata pairwisebalancingstep.then,e(wt)isauniformdistributionatanytimet,and whereb=(?)2+2ands=+b+b2++bd?. E[tde]=(sbtd+d20+s?btd+d?bd2)N?(t+)d2?20; (29) algorithm.aloadbalancingoperationcomprisesdpairwisebalancingstepsinboththetorus andthemeshstructures.toexaminecloselythedynamicbehaviorofthedimension-exchange algorithminthelevelofpairwiseoperations,weintroduceonemorevariablet0todenotethe Proof.Recallthattistheindexofloadbalancingoperationsinthedimension-exchange indexofpairwisesteps.t=0ifandonlyift0=0,andtindexesthetimeinstancest0thatare integermultipliesofd.then,attimet0thatt0=dt, Wt0=MdWt0?+t0 =MdMd?Wt0?2+Mdt0?+t0 =c=dmcwt0?d+2c=dmct0?d+++mdt0?+t0 wherecj=dmj=mdmd?mc,andd+ =MWt0?d+dXc=(c+ j=dmjt0?d+c); j=dmj=. (30) fromtimet0?dtot0,i.e.,,thetthdimension-exchangebalancingoperation.usingindextinsteadoft0,eq.(30)leadsto Let t=pdc=(c+ j=dmjt0?d+c)bethedistributionofworkloadswhicharegenerated/consumed Wt=MWt?+ =MtW0+tXj=Mt?j t 2 j:

22 Usingthelinearityoftheexpectationoperations,E,weobtainthat E(Wt)=E(MtW0+tXj=Mt?j =MtE(W0)+tXj=Mt?jE( j) =0u+dtu; j) whereuisaunitaryvectorofsizen.itisauniformdistribution.therstpartofthelemma tionofworkloadsthataregenerated/consumedintheroundt.then, isthusproved. Wt=Wt?+ Next,weconsidertheworkloadvarianceattimet,E[tde].Let t:bythedenitionofworkloadvariance,wehavetbetheuniformdistribu- t=pdc=t0?d+c,and E[tde]=E(jjWt?Wtjj2) =E(jjMt+W0?W0jj2+tXi=0E(jjMi =E(jjMWt??Wt?jj2)+E(jj t? t+?i? tjj2) t+?ijj2): ToprovethelemmaregardingE[tde],itsucestoshowthatfor0it, Itcanbeshownbyinduction.WerstconsiderE(jj E(jjMi t+?i? t+?ijj2)=bidsn2?d2: of,wehavethat augmentedinaroundofdimension-exchangepairwisebalancingoperations.bythedenition? jj2).itistheworkloadvariance E(jj t? tjj2)=e(jjdxc=(c+ =dxc=[e(jjc+ j=dmjt0?d+c?t0?d+cjj2)] j=dmjt0?d+c?t0?d+c)jj2) =sn2?d2; =d? Xc=(bcN?)2+(N?)2 wherethesecondstepisduetothefactthatc+ ofc+ dentrandomvariablesforcd,andthethirdstepisduetothefollowingreasons.each componentofcj=dmjforanyc,c<d,isrecursivelyacombinationoftwocomponents j=dmjwithcoecients?and.itcanthusbeinferredthatacomponentofcdmj j=dmjt0?d+c?t0?d+carezeromeanindepen- isacombinationof2d?c+componentsofwithcoecientsasfollows. 22

23 b2 b3 d d?2 d? d dd?2d?2 Combinatorialcoecientsai,where=? d d?22d?d?pa2i b Consequently,fromLemma4.,itfollowsthatE(jjcdMj?jj2)=Nbd?c+2?2. Weproceedbyinductiononi.AssumeE(jjMi d t+?i? t+?ijj2)=bidsn2?d2.we bd tisindependentoftaswell.then,e(jjmi+ thenconsidere(jjmi+ t?i? t?ijj2).sincetiisassumedtobeindependentoftimet, suxoperatorsj=dmjredistributestheworkloadsofmi t?i+jj2).fromtheargumentintheprecedingparagraph,itisknownthatasequenceof t?i? t?ijj2)=e(jj(j=dmj)mi insuchawaythateachofits t?i+? whichconcludestheinductionandprovesthesecondpartofthelemma. ofthetable.consequentlye(jjmi+ componentsbecomesacombinationofits2dcomponentswithcoecientsasinthelastrow Fromthelemma,itisevidentthatE[tde]isminimizedat==2overallpossiblechoices t?i? t?ijj2)=bdbidsn2?d2=bid+dsn2?d2, ofthedimension-exchangeparameter.thus,wehave models.noticethatlemma5.2holdsundertheassumptionthattheworkloadgeneration/consumptionratiostiineachpairwisebalancingstepofaroundofdimension-exchange operationhasthesamestatisticalcharacteristicsasthoseinadiusionoperation.itistherefore fairtocomparee[tdf]ofeq.(27)withe[tde]ofeq.(29).considertheall-portcommunication model.substituting=d+forine[tdf]and=2forine[tde],weobtain WefurthercompareE[tade]withE[tadf]inbothone-portandall-portcommunication E[tade]E[tode]: (3) E[tade]=(2? E[tadf]=d+ d[?(d 2d?)?=2t+ d+)t+]n2?(t+)2+?=2dn2?(t+)d2+(2? (d+)t+20?20 Itcanbeeasilyveriedthatthecoecientof20inE[ade]issmallerthanthatinE[tadf],and 2d?) thatthecoecientof2ine[tade]issmallerthanthatine[tadf]whentn=d.hence,for 2td+d20?20: tn=d, interestinpractice. SinceNdinthemeshandthetorus,theaboverelationshipholdsforanytimeinstantof E[tade]E[tadf]: processorinasinglediusionoperationisexpectedtobedwithvarianced2.then,e[tdf] ofeq.(27)becomes Inthecaseoftheone-portcommunicationmodel,theworkloadgenerated/consumedbya (at+20+?at+?ad2)n?(t+)d2?20: 23

24 Clearly,E[tade]E[tadf]atanytimet.Conclusively,weobtainthefollowingtheorem. Theorem5.2Supposeprocessorsarerunningsynchronousdimension-exchangeanddiusion loadbalancingprocessesunderassumption2..then,e[tade]e[tode],e[tadf]e[todf], ande[tade]e[tadf]inbothone-portandall-portcommunicationmodels. Intheprecedingtwosections,weexploredanumberofrelationshipsbetweenthedimension- Numericalexperiments 6exchangeandthediusionmethodswithrespecttotheirecienciesandbalancingqualities. Inordertoobtainanideaofthemagnitudeoftheirdierences,weconductedastatistical networksandusingsyntheticworkloaddistributions.theexperimentalresultsalsoserveto simulationoftheseloadbalancingalgorithmsonvarioustopologiesandsizesofcommunication verifythetheoreticalresults. inastaticworkloadsituation,asimulationofasynchronousloadbalancinginthedynamic situation,andasimulationofsynchronousloadbalancinginthedynamicsituation.ineach simulation,theinitialworkloaddistributionwisassumedtobearandomvector,eachelement Theexperimentincludesthreeparts.Theyareasimulationofsynchronousloadbalancing workloaddistributionsanddierentworkloadgenerationratios.wealsoassumethattheunderlyingsystemimplementstheall-portcommunicationmodelsothatadimension-exchange wofwhichisdrawnindependentlyfromanidenticaluniformdistributionin[0;000].each datapointobtainedintheexperimentistheaverageof20runs,usingdierentrandominitial balancingoperationtakes2ndiusionoperationsinann-dmeshortorus.adiusionoperationistakenasabasictimestepinaloadbalancingprocess. communicationsteps,denotedbyt,necessaryforarrivingatagloballybalancedstate.inthe simulation,wedenetheglobalbalancedstatetobethestateinwhichthesystemworkload varianceislessthanorequaltoone.figure4andfigure5plotthesimulationresultsfrom Inthesimulationofstaticsynchronousloadbalancingprocesses,wemeasurethenumberof dierentloadbalancingalgorithmsexecutedintheringofsizes(n)varyingfrom2to28 nodesandinthe2-dmeshofsizesvaryingfrom22to3232,respectively.thesetwo guresclearlyindicatethatthedimension-exchangemethodoutperformsthediusionmethod acceleratethedimension-exchangeloadbalancingprocesssignicantly.intheringof64nodes, evenintheall-portcommunicationmodel.inparticular,weseethattheodealgorithmdoes forexample,tode=98withtheodealgorithmwhiletade'todf=305andtadf=684with theothers.itsimprovementovertheadealgorithmreachesashighas92:5%.infigure5, balancingprocessina64-ary2-cubeonlyrequiresabout96communicationstepsforarriving observationwasprovedtobetrueinboththemeshandthetorusin[9].thus,anodeload wealsoseethatthenumberofcommunicationstepstina2-dmeshisdependentonlyon thesizeofitslargerdimensionandisinsensitivetothesizeofitssmallerdimension.this ataglobalbalancedstate.itreallyputsforththeodealgorithmasapracticalmethodfor dynamicglobalbalancinginrealmulticomputers. 24

25 8 6 ADE ODE ADF ODF 4 2 log2(t) Figure4:Thenumberofcommunicationstepsnecessaryforreachingagloballybalancedstate 4 duringastaticsynchronousloadbalancingprocessintheringofsizesvaryingfrom4to28 2 nodes log2(n) 8 6 ADE ODE ADF ODF 4 2 log2(t) Figure5:Thenumberofcommunicationstepsnecessaryforreachingagloballybalancedstate 4 duringastaticsynchronousloadbalancingprocessinthe2-dmeshofsizesvaryingfrom22 to x2 4x4 8x4 8x8 6x8 6x6 32x8 32x6 32x32 25

26 Figure6thesystemworkloadvarianceintherst00stepsofvariousloadbalancingprocesses intheringof32nodes.thegureillustratesthattheodealgorithmpullsdownthesystem Furthermore,inordertoexaminetheeectsofasingleloadbalancingoperation,weplotin ADE ODE ADF ODF Figure6:Reductionoftheworkloadvarianceduringastaticsynchronousloadbalancing 200 workloadvariancesharplyalthoughitsinitialreductionratioseemstobenotassatisfactory processintheringof32nodes intheirreductionratios.thissaysthatboththeodeandtheodfalgorithmsmaynot outperformtheirlocalaveragebalancingcounterpartsintheshortterm. asthatoftheadealgorithm.theodfandtheadfalgorithmshavethesamerelationship policysuchthatonceaprocessor'sworkloaddropsorrisesbeyondapairofpresetbounds, processorateachtimestepis00withthevarianceof30andtheconsumptionratioisa constant00.inthesimulationofasynchronousloadbalancing,weuseasimpleinvocation Inthedynamicsituation,weassumethattheexpectedworkloadgenerationratioofa pairofthresholdsdeterminethedegreeofasynchronismofaloadbalancingprocess.suppose wunderloadandwoverload,theprocessorwouldactivatealoadbalancingoperation.evidently,the wunderloadandwoverloadaresymmetricwithrespecttotheexpectedworkloadofaprocessor E(w)=500atanytime,itfollowsthatwunderload=500?range=2andwoverload=500+ range=2.figures7and8plotthesystemworkloadvariancesresultingfromdierentload E(w).Wethenmeasuretherangebetweenwunderloadandwoverloadbyanindexrange.Since balancingalgorithmsinaringof64nodesandameshofsize66forthecaseofrange=600. tunedalgorithmsforglobalsynchronousloadbalancing,donotgainsignicantbenetsin workloadvariancemorerapidlythanthediusionmethodandkeepsitboundedatamuch lowerlevel.itcanalsobeobservedthatboththeodeandtheodfalgorithms,theoptimally Fromthesetwogures,itcanbeseenthattheADEalgorithmreducestheinitialsystem asynchronousimplementations. 26

27 ADE ODE ADF ODF Figure7:Changeoftheworkloadvarianceintherst200stepsofadynamicasynchronous loadbalancingprocessintheringofsize ADE ODE ADF ODF Figure8:Changeoftheworkloadvarianceintherst200stepsofadynamicasynchronous loadbalancingprocessinthemeshofsize

28 ADE ODE ADF ODF Figure9:Changeoftheworkloadvariancesduringadynamicsynchronousloadbalancing processina66torus ADE ODE ADF ODF conductedintherstexperiment,anditsresultsinaringof32nodesarereportedinfigure6. simultaneously.thesimulationofsynchronousimplementationsinthestaticsituationwas mentationsinwhichrangeissettozerosothatallprocessorsparticipateinloadbalancing Synchronousimplementationsofloadbalancingarespecialcasesofasynchronousimple- Figures9and0presentthesimulationresultsofdynamicsynchronousimplementationsinthe 66torusandthe66mesh.InagreementwiththendingsfromFigure6,Figures9 000 Figure0:Changeofthesystemworkloadvarianceduringadynamicsynchronousloadbalancingprocessina66mesh 500 and0showthatthesuperiorityofthedimension-exchangemethodoverthediusionmethod

29 holdsunderthesynchronousinvocationpoliciesaswell,andthattheadealgorithmhasan 7advantageoverthediusionmethodinbothshortandlongterms. algorithms,thedimension-exchange(de)andthediusion(df)methods,withrespectto Inthispaper,wemadeacomparisonbetweentwoclassesofnearestneighborloadbalancing Conclusions theireciencyindrivinganyinitialworkloaddistributiontoauniformdistributionandtheir abilityincontrollingthegrowthofthevarianceamongtheprocessors'workloads.wefocused ontheirfourinstances,theade,theode,theadfandtheodf,whicharethemost synchronous/asynchronousinvocationpoliciesandstatic/dynamicrandomworkloadbehaviors. commonversionsinpractice.thecomparisonwasmadecomprehensivelyinbothone-port andall-portcommunicationmodelswithconsiderationofvariousimplementationstrategies: thataisapproximatelyequivalenttobinperformance.then,ourcomparativeresultscanbe summarizedasintables3and4. Let\ab"denotetherelationshipthataoutperformsb,and\ab"therelationship andn-dtori. Table3:Summaryofcomparativeresultsintheone-portcommunicationmodelinn-Dmeshes Synchronous ODEADEODFADF Staticloadbalancing Dynamicloadbalancing ADEfADF;ODFg ADEODE ADFODF ADEADF Asynchronous ADFODFincasen= ODFADFincasen2 sameasleft Table4:Summaryofcomparativeresultsinall-portcommunicationmodelinn-Dmeshesand n-dtori. Synchronous ODEADEODFADF Staticloadbalancing Dynamicloadbalancing fadf;odfgadeode ADEODE ADFODF ADEADF Asynchronous ADFODFincasen= ODFADFincasen2 sameasleft 29

30 besttosynchronousimplementationinthestaticsituation.wealsorevealedthesuperiority ofthedimension-exchangemethodinsynchronousloadbalancingevenintheall-portcommunicationmodel.thestrengthofthediusionmethodisinasynchronousimplementationin methodintheone-portcommunicationmodel.inparticular,theodealgorithmlendsitself Specically,weshowedthatthedimension-exchangemethodoutperformsthediusion theall-portcommunicationmodel.theodfalgorithmperformsbestinthatcase. algorithms,butalsooerspracticalguidelinestosystemdevelopersindesigningloadbalancing architecturesforvariousparallelcomputationalparadigms.weappliedboththediusionand thedimension-exchangemethodsindistributedbranch-and-boundcomputations,andpartly Thecomparativestudynotonlyprovidesaninsightintonearestneighborloadbalancing intheplatformsofparsytecgcpp(powerpc-based)andparsytecgcel(transputer-based) veriedourcomparativeresultsinbothstaticanddynamicasynchronousimplementations multicomputers[7].wealsoevaluatedtheirsynchronousperformancesinrealapplicationsin periodicre-mappingofdataparallelcomputationsin[9]. ThisworkissupportedinpartbyNSFMIP andtheDFG-Forschergruppe\Eziente Acknowledgments NutzungmassivparallelerSystems".WearegratefultoH.L.Xieforhiscarefulproofreading References andtheanonymousrefereesfortheirvaluablecomments. []I.Ahmad,A.Ghafoor,andKMehrotra.Performancepredictionfordistributedload [2]L.V.Kale.Comparingtheperformanceoftwodynamicloaddistributionmethods.In balancingonmulticomputersystems.inproceedingsofsupercomputing'99,pages830{ 839(99). [3]V.Kumar,A.Y.Grama,andN.R.Vempaty.Scalableloadbalancingtechniquesfor ProceedingsofInternationalConferenceonParallelProcessing,pages8{2(988). [4]M.Willebeek-LeMairandA.P.Reeves.Strategiesfordynamicloadbalancingonhighly parallelcomputers.journalofparallelanddistributedcomputing,22():60{79(994). [5]C.-Z.XuandF.C.M.Lau.Analysisofthegeneralizeddimensionexchangemethodfor parallelcomputers.ieeetransactionsonparallelanddistributedsystems,4(9):979{993 (993). [6]C.-Z.XuandF.C.M.Lau.Iterativedynamicloadbalancinginmulticomputers.Journal dynamicloadbalancing.journalofparallelanddistributedcomputing,6(4):385{393 (992). ofoperationalresearchsociety,45(7):786{796(994). 30

31 [7]D.P.BertsekasandJ.N.Tsitsiklis.Parallelanddistributedcomputation:Numerical [8]G.Cybenko.Loadbalancingfordistributedmemorymultiprocessors.JournalofParallel methods.prentice-hallinc.(989). [9]C.-Z.XuandF.C.M.Lau.Thegeneralizeddimensionexchangemethodforloadbalancing anddistributedcomputing,7:279{30(989). [0]S.L.JohnssonandC.-T.Ho.Spanninggraphsforoptimumbroadcastingandpersonalized ink-aryn-cubesandvariants.journalofparallelanddistributedcomputing,24():72{85 (995). []D.W.Krumme,G.Cybenko,andK.N.Venkataraman.Gossipinginminimaltime.SIAM communicationinhypercubes.ieeetransactionsoncomputers,38(9):249{268(989). [2]J.B.Boillat.Loadbalancingandpoissonequationinagraph.Concurrency:Practice JournalonComputing,2():{39(992). [3]J.-W.Hong,X.-N.Tan,andM.Chen.Fromlocaltoglobal:ananalysisofnearestneighbor andexperience,2(4):289{33(990). [4]X.-S.QianandQ.Yang.Loadbalancingongeneralizedhypercubeandmeshmultiprocessorswithlal.InProceedingsofthInternationalConferenceonDistributedComputing balancingonhypercube.inproceedingsofacm{sigmetrics,pages73{82(988). [5]C.-Z.XuandF.C.M.Lau.Optimalparametersforloadbalancingwiththediusion methodinmeshnetworks.parallelprocessingletters,4(2):39{47(994). Systems,pages402{409(99). [6]S.H.Hosseini,B.Litow,M.Malkawi,J.Mcpherson,andK.Vairavan.Analysisofagraph [7]C.-Z.Xu,S.Tschoeke,andB.Monien.Performanceevaluationofloaddistributionstrategiesinparallelbranch-and-boundcomputations.Technicalreport,Dept.ofElectricaland coloringbaseddistributedloadbalancingalgorithm.journalofparallelanddistributed Computing,0:60{66(990). [8]R.Diekmann,D.Meyer,andB.Monien.ParalleldecompositionofunstructuredFEMmeshes.Technicalreport,Dept.ofMathematicsandComputerScience,Universityof Paderborn,Germany(995). ComputerEngg.,WayneStateUniversity(995). [9]C.-Z.XuandF.C.M.Lau.Decentralizedremappingofdata-parallelcomputationswith [20]J.Song.Apartiallyasynchronousanditerativealgorithmfordistributedloadbalancing. formancecomputingconference,pages44{42.ieeecomputersocietypress(994). thegeneralizeddimensionexchangemethod.inproceedingsof994scalablehighper- ParallelComputing,20(6):853{868(994). 3

32 [2]R.LulingandB.Monien.Adynamicdistributedloadbalancingalgorithmwithprovable [22]W.J.Dally.Performanceanalysisofk-aryn-cubeinterconnectionnetworks.IEEETransactionsonComputers,39(6):775{785(990). goodperformance.inproceedingsof5thacmsymposiumonparallelalgorithmsand Architectures,pages64{72(993). [23]L.M.NiandP.K.McKinley.Asurveyofwormholeroutingtechniquesindirectnetworks. [24]G.RamanathanandJ.Oren.Surveyofcommercialparallelmachines.ACMComputer ArchitectureNews,2(3):3{33(993). IEEEComputer,26:62{76(993). [26]S.Ranka,Y.Won,andS.Sahni.Programmingahypercubemulticomputer.IEEE [25]D.M.NicolandJ.H.Saltz.Dynamicremappingofparallelcomputationswithvarying Software,5:69{77(988). resourcedemands.ieeetransactionsoncomputers,37(9):073{087(988). [27]Y.ShihandJ.Fier.Hypercubesystemsandkeyapplications.InK.HwangandD.Degroot,editors,ParallelProcessingforSupercomputersandArticalIntelligence,pages 203{243.McGraw-HillPublishingCo.(989). [28]S.FioriniandR.J.Wilson.Edge-coloringofgraphs.InL.W.BeinekeandR.J.Wilson, [29]B.GhoshandS.Muthukrishnan.Dynamicloadbalancingindistributednetworksby randommatchings.inproceedingsof6thacmsymposiumonparallelalgorithmsand editors,selectedtopicsingraphtheory,pages03{25.academicpress(978). [30]A.N.Choudhary,B.Narahari,andR.Krishnamurti.Anecientheuristicschemefor dynamicremappingofparallelcomputations.parallelcomputing,9:62{632(993). Architectures(994). [3]J.DeKeyserandD.Roose.Loadbalancingdataparallelprogramsondistributedmemory computers.parallelcomputing,9:99{29(993). 32

( ) = ( ) = {,,, } β ( ), < 1 ( ) + ( ) = ( ) + ( )

( ) = ( ) = {,,, } β ( ), < 1 ( ) + ( ) = ( ) + ( ) { } ( ) = ( ) = {,,, } ( ) β ( ), < 1 ( ) + ( ) = ( ) + ( ) max, ( ) [ ( )] + ( ) [ ( )], [ ( )] [ ( )] = =, ( ) = ( ) = 0 ( ) = ( ) ( ) ( ) =, ( ), ( ) =, ( ), ( ). ln ( ) = ln ( ). + 1 ( ) = ( ) Ω[ (

More information

Topological Properties

Topological Properties Advanced Computer Architecture Topological Properties Routing Distance: Number of links on route Node degree: Number of channels per node Network diameter: Longest minimum routing distance between any

More information

Performance Comparison of Dynamic Load-Balancing Strategies for Distributed Computing

Performance Comparison of Dynamic Load-Balancing Strategies for Distributed Computing Performance Comparison of Dynamic Load-Balancing Strategies for Distributed Computing A. Cortés, A. Ripoll, M.A. Senar and E. Luque Computer Architecture and Operating Systems Group Universitat Autònoma

More information

ToappearinJ.ofParallelandDistributedProcessing. TheGeneralizedDimensionExchangeMethodforLoad Balancingink-aryn-cubesandVariants

ToappearinJ.ofParallelandDistributedProcessing. TheGeneralizedDimensionExchangeMethodforLoad Balancingink-aryn-cubesandVariants ToappearinJ.ofParallelandDistributedProcessing TheGeneralizedDimensionExchangeMethodforLoad Balancingink-aryn-cubesandVariants DepartmentofComputerScience,TheUniversityofHongKong,HongKong DepartmentofComputerScience,ShantouUniversity,P.R.China

More information

CONTROLLER INFORMATION SHEET

CONTROLLER INFORMATION SHEET CONTROLLER INFORMATION SHEET Maple Model(s) Graphic HMIs PLC or Controller Siemens LOGO! P/N: 1036-0230 Rev. 00 Date: 01/11/2016 Summary Maple Systems Graphic HMIs communicate with the Siemens LOGO! controller

More information

Panasonic FP. HMI Setting: Device Address:

Panasonic FP. HMI Setting: Device Address: Panasonic FP Supported Series: NAIS (Matsushita) FP series include FP-X, FP-Σ, FP0, FP1, FP2, FP2SH, FP10SH and FP3 Ethernet support FP-X with AFPX-COM5. Website:http://pewa.panasonic.com/ HMI Setting:

More information

thek-aryn-cubestructure. 1

thek-aryn-cubestructure. 1 DDE:AModiedDimensionExchangeMethod forloadbalancingink-aryn-cubes StateUniversityofNewYorkatBualo DepartmentofComputerScience Min-YouWuandWeiShu algorithmforthehypercubestructure.ithasbeengeneralizedtok-aryn-cubes.however,the

More information

Currency Options (2): Hedging and Valuation

Currency Options (2): Hedging and Valuation Overview Chapter 9 (2): Hedging and Overview Overview The Replication Approach The Hedging Approach The Risk-adjusted Probabilities Notation Discussion Binomial Option Pricing Backward Pricing, Dynamic

More information

Themethodofmovingcurvesandmovingsurfacesisanew,eectivetoolfor Abstract

Themethodofmovingcurvesandmovingsurfacesisanew,eectivetoolfor Abstract OnaRelationshipbetweentheMovingLineand MovingConicCoecientMatrices DepartmentofComputerScience Houston,Texas77005 [email protected] RiceUniversity MingZhang DepartmentofInformationSystemsandComputerScience

More information

estadium Project Lab 8: Wireless Mesh Network Setup with DD WRT

estadium Project Lab 8: Wireless Mesh Network Setup with DD WRT estadium Project Lab 8: Wireless Mesh Network Setup with DD WRT Objectives To become familiar with wireless mesh networks and show set up a wireless mesh network test bed using the DD WRT firmware. We

More information

Implementing and Managing Windows Server 2008 Clustering

Implementing and Managing Windows Server 2008 Clustering Implementing and Managing Windows Server 2008 Clustering Course Number: 6423A Course Length: 3 Days Course Overview This instructor-led course explores Windows Server 2008 clustering and provides students

More information

ParFUM: A Parallel Framework for Unstructured Meshes. Aaron Becker, Isaac Dooley, Terry Wilmarth, Sayantan Chakravorty Charm++ Workshop 2008

ParFUM: A Parallel Framework for Unstructured Meshes. Aaron Becker, Isaac Dooley, Terry Wilmarth, Sayantan Chakravorty Charm++ Workshop 2008 ParFUM: A Parallel Framework for Unstructured Meshes Aaron Becker, Isaac Dooley, Terry Wilmarth, Sayantan Chakravorty Charm++ Workshop 2008 What is ParFUM? A framework for writing parallel finite element

More information

RAID. Storage-centric computing, cloud computing. Benefits:

RAID. Storage-centric computing, cloud computing. Benefits: RAID Storage-centric computing, cloud computing. Benefits: Improved reliability (via error correcting code, redundancy). Improved performance (via redundancy). Independent disks. RAID Level 0 Provides

More information

(Master Slave Mode) 28-1. This chapter explains how to connect multiple HMIs.

(Master Slave Mode) 28-1. This chapter explains how to connect multiple HMIs. 28-1 28.Multi-HMI Communication (Master Slave Mode) This chapter explains how to connect multiple HMIs. 28.1. Overview... 28-2 28.2. Steps to Create a Project of Master HMI... 28-2 28.3. Steps to Create

More information

The integrating factor method (Sect. 2.1).

The integrating factor method (Sect. 2.1). The integrating factor method (Sect. 2.1). Overview of differential equations. Linear Ordinary Differential Equations. The integrating factor method. Constant coefficients. The Initial Value Problem. Variable

More information

Using AD fields in Policy Patrol

Using AD fields in Policy Patrol Policy Patrol 9 technical documentation May 20, 2013 in Policy Patrol This document describes how to enter additional Active Directory merge fields in Policy Patrol and how to convert AD fields into a

More information

Tutorial: Structural Models of the Firm

Tutorial: Structural Models of the Firm Tutorial: Structural Models of the Firm Peter Ritchken Case Western Reserve University February 16, 2015 Peter Ritchken, Case Western Reserve University Tutorial: Structural Models of the Firm 1/61 Tutorial:

More information

Continual Reassessment Method

Continual Reassessment Method Continual Reassessment Method Adrian Mander MRC Biostatistics Unit Hub for Trials Methodology Research, Cambridge Sep 2011 Adrian Mander Sep 2011 1/17 Outline Introduction to oncology phase I trials Describe

More information

Interconnection Networks

Interconnection Networks Advanced Computer Architecture (0630561) Lecture 15 Interconnection Networks Prof. Kasim M. Al-Aubidy Computer Eng. Dept. Interconnection Networks: Multiprocessors INs can be classified based on: 1. Mode

More information

L 2 : x = s + 1, y = s, z = 4s + 4. 3. Suppose that C has coordinates (x, y, z). Then from the vector equality AC = BD, one has

L 2 : x = s + 1, y = s, z = 4s + 4. 3. Suppose that C has coordinates (x, y, z). Then from the vector equality AC = BD, one has The line L through the points A and B is parallel to the vector AB = 3, 2, and has parametric equations x = 3t + 2, y = 2t +, z = t Therefore, the intersection point of the line with the plane should satisfy:

More information

New Product Hotline - 800-336-3575

New Product Hotline - 800-336-3575 Dia. M12-1.25 (continued) 6-242 M12-1.25 12.56mm 37.5mm 11.5mm 6-375 M12-1.25 14.25mm 41.5mm 14mm 6-263 M12-1.25 12.83mm 32.5mm 7mm 6-320 M12-1.25 12.83mm 41.5mm 14mm 6-322 M12-1.25 14.27mm 30.5mm 6mm

More information

ACTS 4302 SOLUTION TO MIDTERM EXAM Derivatives Markets, Chapters 9, 10, 11, 12, 18. October 21, 2010 (Thurs)

ACTS 4302 SOLUTION TO MIDTERM EXAM Derivatives Markets, Chapters 9, 10, 11, 12, 18. October 21, 2010 (Thurs) Problem ACTS 4302 SOLUTION TO MIDTERM EXAM Derivatives Markets, Chapters 9, 0,, 2, 8. October 2, 200 (Thurs) (i) The current exchange rate is 0.0$/. (ii) A four-year dollar-denominated European put option

More information

Lecture 2 Parallel Programming Platforms

Lecture 2 Parallel Programming Platforms Lecture 2 Parallel Programming Platforms Flynn s Taxonomy In 1966, Michael Flynn classified systems according to numbers of instruction streams and the number of data stream. Data stream Single Multiple

More information

Jorge Cruz Lopez - Bus 316: Derivative Securities. Week 9. Binomial Trees : Hull, Ch. 12.

Jorge Cruz Lopez - Bus 316: Derivative Securities. Week 9. Binomial Trees : Hull, Ch. 12. Week 9 Binomial Trees : Hull, Ch. 12. 1 Binomial Trees Objective: To explain how the binomial model can be used to price options. 2 Binomial Trees 1. Introduction. 2. One Step Binomial Model. 3. Risk Neutral

More information

Interconnection Network Design

Interconnection Network Design Interconnection Network Design Vida Vukašinović 1 Introduction Parallel computer networks are interesting topic, but they are also difficult to understand in an overall sense. The topological structure

More information

Interconnection Network

Interconnection Network Interconnection Network Recap: Generic Parallel Architecture A generic modern multiprocessor Network Mem Communication assist (CA) $ P Node: processor(s), memory system, plus communication assist Network

More information

Copyright 2011 - bizagi

Copyright 2011 - bizagi Copyright 2011 - bizagi 1. Process Automation with bizagi... 3 Description... 3 Objectives... 3 Target Audience Profile... 4 Duration... 4 2. Part I Basic concepts to build a bizagi solution... 5 Description...

More information

Version of Barcode Toolbox adds support for Adobe Illustrator CS

Version of Barcode Toolbox adds support for Adobe Illustrator CS Introduction Traditionally the printing industry has used manual design paste up and stripping for the implementation of barcode symbols but today the computer is used more often to generate barcode symbols

More information

Features and Benefits

Features and Benefits Refrigeration 2 eatures and Benefits Based on IR32 Parameters Real Time lock ( RT ) Model Specific (defrost based on real time 7 day 24 hour) HAP (Hazard Analysis and ritical ontrol Point) Alarm Available

More information

Example: Determine the power supplied by each of the sources, independent and dependent, in this circuit:

Example: Determine the power supplied by each of the sources, independent and dependent, in this circuit: Example: Determine the power supplied by each of the sources, independent and dependent, in this circuit: Solution: We ll begin by choosing the bottom node to be the reference node. Next we ll label the

More information

ES250: Electrical Science. HW7: Energy Storage Elements

ES250: Electrical Science. HW7: Energy Storage Elements ES250: Electrical Science HW7: Energy Storage Elements Introduction This chapter introduces two more circuit elements, the capacitor and the inductor whose elements laws involve integration or differentiation;

More information

Homework 2 Solutions

Homework 2 Solutions Homework Solutions Igor Yanovsky Math 5B TA Section 5.3, Problem b: Use Taylor s method of order two to approximate the solution for the following initial-value problem: y = + t y, t 3, y =, with h = 0.5.

More information

Nominal rates of interest and discount

Nominal rates of interest and discount 1 Nominal rates of interest and discount i (m) : The nominal rate of interest payable m times per period, where m is a positive integer > 1. By a nominal rate of interest i (m), we mean a rate payable

More information

Experiences of numerical simulations on a PC cluster Antti Vanne December 11, 2002

Experiences of numerical simulations on a PC cluster Antti Vanne December 11, 2002 xperiences of numerical simulations on a P cluster xperiences of numerical simulations on a P cluster ecember xperiences of numerical simulations on a P cluster Introduction eowulf concept Using commodity

More information

Stirling s formula, n-spheres and the Gamma Function

Stirling s formula, n-spheres and the Gamma Function Stirling s formula, n-spheres and the Gamma Function We start by noticing that and hence x n e x dx lim a 1 ( 1 n n a n n! e ax dx lim a 1 ( 1 n n a n a 1 x n e x dx (1 Let us make a remark in passing.

More information

Electricity & Gas Prices in Ireland. 1 st Semester 2013

Electricity & Gas Prices in Ireland. 1 st Semester 2013 Electricity & Gas Prices in Ireland 1 st Semester 2013 Overview Energy Prices to Business Highlights Trends and EU comparisons Energy Prices to Households Highlights Trends and EU comparisons Global Oil

More information

Load Balancing. Load Balancing 1 / 24

Load Balancing. Load Balancing 1 / 24 Load Balancing Backtracking, branch & bound and alpha-beta pruning: how to assign work to idle processes without much communication? Additionally for alpha-beta pruning: implementing the young-brothers-wait

More information

This chapter includes installation instructions and limitations for Antivirus products on client computers and loggers.

This chapter includes installation instructions and limitations for Antivirus products on client computers and loggers. : tiiu Ti t iud itti ituti imitti f tiiu dut it mut gg NT: T ifmti i ti t f t ftw i y I dditi, utm, bui t, i mut ify tt t Lgg mt t miimum dw quimt dfi by t tid ty ftw Ctt G tiiu 9 f NIC Itti gmt Tid ty

More information

Sample Solutions for Assignment 2.

Sample Solutions for Assignment 2. AMath 383, Autumn 01 Sample Solutions for Assignment. Reading: Chs. -3. 1. Exercise 4 of Chapter. Please note that there is a typo in the formula in part c: An exponent of 1 is missing. It should say 4

More information

Aperiodic Task Scheduling

Aperiodic Task Scheduling Aperiodic Task Scheduling Jian-Jia Chen (slides are based on Peter Marwedel) TU Dortmund, Informatik 12 Germany Springer, 2010 2014 年 11 月 19 日 These slides use Microsoft clip arts. Microsoft copyright

More information

How to Fix Mail-Merge Number Formatting in Word 2010

How to Fix Mail-Merge Number Formatting in Word 2010 How to Fix Mail-Merge Number Formatting in Word 2010 By Rich Malloy, Tech Help Today, June 2012, updated Nov. 2012 When 93.90 turns into 93.90000000006, there are no less than three ways to put things

More information

Lesson 14 14 Outline Outline

Lesson 14 14 Outline Outline Lesson 14 Confidence Intervals of Odds Ratio and Relative Risk Lesson 14 Outline Lesson 14 covers Confidence Interval of an Odds Ratio Review of Odds Ratio Sampling distribution of OR on natural log scale

More information

2 Basic Concepts. Contents

2 Basic Concepts. Contents 2. Basic Concepts Contents 2 Basic Concepts a. Link configuration b. Topology c. Transmission mode d. Classes of networks 1 a. Link Configuration Data links A direct data link is one that establishes a

More information

Components: Interconnect Page 1 of 18

Components: Interconnect Page 1 of 18 Components: Interconnect Page 1 of 18 PE to PE interconnect: The most expensive supercomputer component Possible implementations: FULL INTERCONNECTION: The ideal Usually not attainable Each PE has a direct

More information

Math 22B, Homework #8 1. y 5y + 6y = 2e t

Math 22B, Homework #8 1. y 5y + 6y = 2e t Math 22B, Homework #8 3.7 Problem # We find a particular olution of the ODE y 5y + 6y 2e t uing the method of variation of parameter and then verify the olution uing the method of undetermined coefficient.

More information

The Goldberg Rao Algorithm for the Maximum Flow Problem

The Goldberg Rao Algorithm for the Maximum Flow Problem The Goldberg Rao Algorithm for the Maximum Flow Problem COS 528 class notes October 18, 2006 Scribe: Dávid Papp Main idea: use of the blocking flow paradigm to achieve essentially O(min{m 2/3, n 1/2 }

More information

1099 and W2 Tax Form Tips and Instructions for 2013 (Effective January 1, 2014)

1099 and W2 Tax Form Tips and Instructions for 2013 (Effective January 1, 2014) 1099 and W2 Tax Form Tips and Instructions for 2013 (Effective January 1, 2014) This document outlines the various forms that can be used to print W-2s in the Infinity POWER Payroll module and 1099s in

More information

1.(6pts) Find symmetric equations of the line L passing through the point (2, 5, 1) and perpendicular to the plane x + 3y z = 9.

1.(6pts) Find symmetric equations of the line L passing through the point (2, 5, 1) and perpendicular to the plane x + 3y z = 9. .(6pts Find symmetric equations of the line L passing through the point (, 5, and perpendicular to the plane x + 3y z = 9. (a x = y + 5 3 = z (b x (c (x = ( 5(y 3 = z + (d x (e (x + 3(y 3 (z = 9 = y 3

More information

How To - Implement Clientless Single Sign On Authentication in Single Active Directory Domain Controller Environment

How To - Implement Clientless Single Sign On Authentication in Single Active Directory Domain Controller Environment How To - Implement Clientless Single Sign On Authentication in Single Active Directory Domain Controller Environment How To - Implement Clientless Single Sign On Authentication with Active Directory Applicable

More information

. MEDIUM SPEED OPERATION - 8MHz (typ.) @ . MULTI-PACKAGE PARALLEL CLOCKING FOR HCC4029B HCF4029B PRESETTABLE UP/DOWN COUNTER BINARY OR BCD DECADE

. MEDIUM SPEED OPERATION - 8MHz (typ.) @ . MULTI-PACKAGE PARALLEL CLOCKING FOR HCC4029B HCF4029B PRESETTABLE UP/DOWN COUNTER BINARY OR BCD DECADE HCC4029B HCF4029B PRESETTABLE UP/DOWN COUNTER BINARY OR BCD DECADE. MEDIUM SPEED OPERATION - 8MHz (typ.) @ CL = 50pF AND DD-SS = 10. MULTI-PACKAGE PARALLEL CLOCKING FOR SYNCHRONOUS HIGH SPEED OUTPUT RES-

More information

Intrusion Log Sharing University of Wisconsin-Madison

Intrusion Log Sharing University of Wisconsin-Madison Intrusion Log Sharing University of Wisconsin-Madison John Bethencourt ([email protected]) Jason Franklin ([email protected]) Mary Vernon ([email protected]) 1 Talk Outline Background: Blacklists,

More information

Exam MFE/3F Sample Questions and Solutions #1 to #76

Exam MFE/3F Sample Questions and Solutions #1 to #76 Exam MFE/3F Sample Questions and Solutions #1 to #76 In this version, standard normal distribution values are obtained by using the Cumulative Normal Distribution Calculator and Inverse CDF Calculator

More information

Architectural Level Power Consumption of Network on Chip. Presenter: YUAN Zheng

Architectural Level Power Consumption of Network on Chip. Presenter: YUAN Zheng Architectural Level Power Consumption of Network Presenter: YUAN Zheng Why Architectural Low Power Design? High-speed and large volume communication among different parts on a chip Problem: Power consumption

More information

Technical Bulletin. Teledyne PDS Clock Synchronization Considerations. Version 1.2

Technical Bulletin. Teledyne PDS Clock Synchronization Considerations. Version 1.2 Teledyne PDS Clock Synchronization Considerations Version 1.2 TELEDYNE RESON B.V. Stuttgartstraat 42-44 3047 AS Rotterdam The Netherlands Tel.: +31 (0)10 245 15 00 www.teledyne-reson.com Dated: 01-05-2015

More information

1 The Black-Scholes model: extensions and hedging

1 The Black-Scholes model: extensions and hedging 1 The Black-Scholes model: extensions and hedging 1.1 Dividends Since we are now in a continuous time framework the dividend paid out at time t (or t ) is given by dd t = D t D t, where as before D denotes

More information

Next Generation Siebel Monitoring: A Real World Customer Experience. An Oracle White Paper June 2010

Next Generation Siebel Monitoring: A Real World Customer Experience. An Oracle White Paper June 2010 Next Generation Siebel Monitoring: A Real World Customer Experience An Oracle White Paper June 2010 Next Generation Siebel Monitoring: A Real World Customer Experience Table of Contents Introduction...

More information

How To - Implement Single Sign On Authentication with Active Directory

How To - Implement Single Sign On Authentication with Active Directory How To - Implement Single Sign On Authentication with Active Directory Applicable to English version of Windows This article describes how to implement single sign on authentication with Active Directory

More information

Attachment "A" - List of HP Inkjet Printers

Attachment A - List of HP Inkjet Printers HP Deskjet 350c Printer HP Deskjet 350cbi Printer HP Deskjet 350cbi Printer w/roller-case HP Deskjet 420 Printer HP Deskjet 420c Printer HP Deskjet 610c Printer HP Deskjet 610cl Printer HP Deskjet 612c

More information

EC3070 FINANCIAL DERIVATIVES

EC3070 FINANCIAL DERIVATIVES BINOMIAL OPTION PRICING MODEL A One-Step Binomial Model The Binomial Option Pricing Model is a simple device that is used for determining the price c τ 0 that should be attributed initially to a call option

More information

ISDN SIGNALLING MODULE SINGLE E1/T1

ISDN SIGNALLING MODULE SINGLE E1/T1 IDN IGNALLING ODUL INGL 1/T1 NODO IDN IGNALLING ODUL INGL 1/T1 (I INGL) is designed for signalling over a single Primary Rate Interface (PRI) of public or private Integrated ervices Digital Networks (IDN)

More information

Answers to Sample Questions on Network Layer

Answers to Sample Questions on Network Layer Answers to Sample Questions on Network Layer ) IP Packets on a certain network can carry a maximum of only 500 bytes in the data portion. An application using TCP/IP on a node on this network generates

More information

Solutions to old Exam 1 problems

Solutions to old Exam 1 problems Solutions to old Exam 1 problems Hi students! I am putting this old version of my review for the first midterm review, place and time to be announced. Check for updates on the web site as to which sections

More information

SOC architecture and design

SOC architecture and design SOC architecture and design system-on-chip (SOC) processors: become components in a system SOC covers many topics processor: pipelined, superscalar, VLIW, array, vector storage: cache, embedded and external

More information

How to Design a Form Report (RTF) Output

How to Design a Form Report (RTF) Output How to Design a Form Report (RTF) Output Applicable to SIMS.net version 7.108 onwards. Permissions required You will need to be a member of any of the following user groups in System Manager to design

More information

Monte Carlo Experiment With Path Dependent Trader Survival Rates

Monte Carlo Experiment With Path Dependent Trader Survival Rates Monte Carlo Experiment With Path Dependent Trader Survival Rates Which Ones Are Preferable, a Cancer Patient's or a Trader's 5-Year Survival Rates? Nassim Nicholas Taleb October 2003 Luck The idea is to

More information

i n g S e c u r it y 3 1B# ; u r w e b a p p li c a tio n s f r o m ha c ke r s w ith t his å ] í d : L : g u id e Scanned by CamScanner

i n g S e c u r it y 3 1B# ; u r w e b a p p li c a tio n s f r o m ha c ke r s w ith t his å ] í d : L : g u id e Scanned by CamScanner í d : r ' " B o m m 1 E x p e r i e n c e L : i i n g S e c u r it y. 1-1B# ; u r w e b a p p li c a tio n s f r o m ha c ke r s w ith t his g u id e å ] - ew i c h P e t e r M u la e n PACKT ' TAÞ$Æo

More information

Big Data: Opportunities and Challenges for Complex Networks

Big Data: Opportunities and Challenges for Complex Networks Big Data: Opportunities and Challenges for Complex Networks Jinhu Lü Academy of Mathematics and Systems Science Chinese Academy of Sciences IWCSN, Vancouver, 13 Dec 2013 1 Thanks To Prof. Ljiljana j Trajkovic

More information

Cray Gemini Interconnect. Technical University of Munich Parallel Programming Class of SS14 Denys Sobchyshak

Cray Gemini Interconnect. Technical University of Munich Parallel Programming Class of SS14 Denys Sobchyshak Cray Gemini Interconnect Technical University of Munich Parallel Programming Class of SS14 Denys Sobchyshak Outline 1. Introduction 2. Overview 3. Architecture 4. Gemini Blocks 5. FMA & BTA 6. Fault tolerance

More information

InHand Device Cloud Service DN 4.0 Quick Start Guide

InHand Device Cloud Service DN 4.0 Quick Start Guide InHand Device Cloud Service DN 4.0 Quick Start Guide Contents 1. Overview... 3 2. Device Manager... 4 2.1 Get an account by self-registration... 4 2.2 Remotely Configure Gateway and Upgrade firmware...

More information

524 Computer Networks

524 Computer Networks 524 Computer Networks Section 1: Introduction to Course Dr. E.C. Kulasekere Sri Lanka Institute of Information Technology - 2005 Course Outline The Aim The course is design to establish the terminology

More information

System Interconnect Architectures. Goals and Analysis. Network Properties and Routing. Terminology - 2. Terminology - 1

System Interconnect Architectures. Goals and Analysis. Network Properties and Routing. Terminology - 2. Terminology - 1 System Interconnect Architectures CSCI 8150 Advanced Computer Architecture Hwang, Chapter 2 Program and Network Properties 2.4 System Interconnect Architectures Direct networks for static connections Indirect

More information

Chapter 4 Multi-Stage Interconnection Networks The general concept of the multi-stage interconnection network, together with its routing properties, have been used in the preceding chapter to describe

More information

Here, we will discuss step-by-step procedure for enabling LDAP Authentication.

Here, we will discuss step-by-step procedure for enabling LDAP Authentication. LDAP Authenticated Web Administration : MailScan 5.x is powered with LDAP Authenticated Web Administration. This gives security enhancement to authenticate users, to check their quarantined and ham emails.

More information

Agenda. Federation using ADFS and Extensibility options. Office 365 Identity overview. Federation and Synchronization

Agenda. Federation using ADFS and Extensibility options. Office 365 Identity overview. Federation and Synchronization Agenda Office 365 Identity overview 1 Federation and Synchronization Federation using ADFS and Extensibility options 2 3 What s New in Azure AD? Cloud Business App - Overview 4 Identity Management is

More information

Hyper Node Torus: A New Interconnection Network for High Speed Packet Processors

Hyper Node Torus: A New Interconnection Network for High Speed Packet Processors 2011 International Symposium on Computer Networks and Distributed Systems (CNDS), February 23-24, 2011 Hyper Node Torus: A New Interconnection Network for High Speed Packet Processors Atefeh Khosravi,

More information

General Theory of Differential Equations Sections 2.8, 3.1-3.2, 4.1

General Theory of Differential Equations Sections 2.8, 3.1-3.2, 4.1 A B I L E N E C H R I S T I A N U N I V E R S I T Y Department of Mathematics General Theory of Differential Equations Sections 2.8, 3.1-3.2, 4.1 Dr. John Ehrke Department of Mathematics Fall 2012 Questions

More information

Fore! Reservations. Integrated Debit Processing

Fore! Reservations. Integrated Debit Processing Fore! 2008 Reservations This document will help explain the process of installing and using the integrated Canadian debit processing in Fore! Reservations 2008. Integrated Debit Processing Table of Contents

More information

Interconnection Networks Programmierung Paralleler und Verteilter Systeme (PPV)

Interconnection Networks Programmierung Paralleler und Verteilter Systeme (PPV) Interconnection Networks Programmierung Paralleler und Verteilter Systeme (PPV) Sommer 2015 Frank Feinbube, M.Sc., Felix Eberhardt, M.Sc., Prof. Dr. Andreas Polze Interconnection Networks 2 SIMD systems

More information

Manual for SOA Exam MLC.

Manual for SOA Exam MLC. Chapter 4. Life Insurance. c 29. Miguel A. Arcones. All rights reserved. Extract from: Arcones Manual for the SOA Exam MLC. Fall 29 Edition. available at http://www.actexmadriver.com/ c 29. Miguel A. Arcones.

More information

Communication Networks. MAP-TELE 2011/12 José Ruela

Communication Networks. MAP-TELE 2011/12 José Ruela Communication Networks MAP-TELE 2011/12 José Ruela Network basic mechanisms Introduction to Communications Networks Communications networks Communications networks are used to transport information (data)

More information

8741A UNIVERSAL PERIPHERAL INTERFACE 8-BIT MICROCOMPUTER

8741A UNIVERSAL PERIPHERAL INTERFACE 8-BIT MICROCOMPUTER UNIVERSAL PERIPHERAL INTERFACE 8-BIT MICROCOMPUTER 8-Bit CPU plus ROM RAM I O Timer and Clock in a Single Package One 8-Bit Status and Two Data Registers for Asynchronous Slave-to- Master Interface DMA

More information

DIGITAL COUNTERS. Q B Q A = 00 initially. Q B Q A = 01 after the first clock pulse.

DIGITAL COUNTERS. Q B Q A = 00 initially. Q B Q A = 01 after the first clock pulse. DIGITAL COUNTERS http://www.tutorialspoint.com/computer_logical_organization/digital_counters.htm Copyright tutorialspoint.com Counter is a sequential circuit. A digital circuit which is used for a counting

More information

LDAP Operation Guide

LDAP Operation Guide LDAP Operation Guide (Lightweight Directory Access Protocol) To find basic information about network and advanced network features of your Brother machine: See the uu Network User's Guide. To download

More information

- Nishad Nerurkar. - Aniket Mhatre

- Nishad Nerurkar. - Aniket Mhatre - Nishad Nerurkar - Aniket Mhatre Single Chip Cloud Computer is a project developed by Intel. It was developed by Intel Lab Bangalore, Intel Lab America and Intel Lab Germany. It is part of a larger project,

More information

Pull versus Push Mechanism in Large Distributed Networks: Closed Form Results

Pull versus Push Mechanism in Large Distributed Networks: Closed Form Results Pull versus Push Mechanism in Large Distributed Networks: Closed Form Results Wouter Minnebo, Benny Van Houdt Dept. Mathematics and Computer Science University of Antwerp - iminds Antwerp, Belgium Wouter

More information

Asynchronous Bypass Channels

Asynchronous Bypass Channels Asynchronous Bypass Channels Improving Performance for Multi-Synchronous NoCs T. Jain, P. Gratz, A. Sprintson, G. Choi, Department of Electrical and Computer Engineering, Texas A&M University, USA Table

More information

IE1204 Digital Design F12: Asynchronous Sequential Circuits (Part 1)

IE1204 Digital Design F12: Asynchronous Sequential Circuits (Part 1) IE1204 Digital Design F12: Asynchronous Sequential Circuits (Part 1) Elena Dubrova KTH / ICT / ES [email protected] BV pp. 584-640 This lecture IE1204 Digital Design, HT14 2 Asynchronous Sequential Machines

More information

Manual for SOA Exam MLC.

Manual for SOA Exam MLC. Chapter 4. Life Insurance. Extract from: Arcones Manual for the SOA Exam MLC. Fall 2009 Edition. available at http://www.actexmadriver.com/ 1/14 Level benefit insurance in the continuous case In this chapter,

More information