You are on page 1of 7

Thespecificnumbersofabovemeasurementsde pendonthebenchmarkschosenandthe compilertechnologyused.

Althoughwefeelthat ourmeasurementsarereasonablyindicativeof theusageofthetwoarchitectures,otherprogra msordifferentcompilersmayyieldslightly differentnumbers.Butwebelievewhatwedoh eretoIntel80386andMIPSR2000primarily displaysthemostcommondifferencesbetweent hetwoarchitecturesbehindthem. 4.Discussions 4.1Whowins? RISCprocessorsgainareputationforhighperfo rmance,andourexperimentsabovedoverify thispoint.However,nowadays,thedifferencebet weenRISCandCISCchipsisgettingsmaller andsmaller.RISCandCISCarchitecturesareb ecomingmoreandmorealike.Manyoftoday's RISCchipssupportjustasmanyinstructionsas yesterday'sCISCchips.ThePowerPC601,for example,supportsmoreinstructionsthanthePen tium.Yetthe601isconsideredaRISCchip, whilethePentiumisdefinitelyCISC.Furthermor etoday'sCISCchipsusemanytechniques formerlyassociatedwithRISCchips.Heremaybe wecansimplysaythatRISCandCISCare growingtoeachotherintheirarchitectureinthe theoreticalpointofview. Inreality,whatcountsishowfastachipcane xecutetheinstructionsitisgivenandhowwelli

The typicalhighendembeddedproductthesedaysincorporatesc omplexsoftware.Intelrapidl yloweredPentiumprices. Makingaprocessorsuccessfulismorethanjust havingthefastestchipavailableatanattractive price.Today.Thisnecessitatesaproducti vesoftware developmentenvironment.pipelining.Inan efforttoacceleratetheintroductionof5th generationprocessorsonthedesktop.In90's.Thi sallhappenedquicklyand .includingmultipleexecution units.Consequentlyhighperformance 486processors-costingmulti-millionsof dollarstodevelopbecameavailableatveryaggressiveprices.yetmostdeal withshrinkingtime-tomarketrequirements.ManyofthenewerRI SCprocessorscannotyetofferatools environmentcomparablewithx86offerings.andfloating-point integration.thusmakesperformancecomparisons almostuselessoutsideofaspecific application.Thereisaproblemofensuringqualitysoft waredevelopmenttoolsareavailable.bothRISCandCI SCmanufacturersaredoingeverythingtogetan edgeonthecompetition.caches.mostnewgen erationsofprocessorsemployamixedbagof architecturalfeatures.forcingdownthe priceofthe486.t runsexistingsoftware.

whowins?Noone wins.CISCba ttlefield.bu tmakesbothextinct.EPICstands .changedthelandscapeoftheRISCvs.Embeddedproductdesignersarego odatidentifyingwherethebestvalueis.[5] ThedebatebetweenRISCandCISCwilllikelyc ontinue.wewanttopointoutthatthebiggestthr eatforCISCandRISCmightnotbeeachother .nomatterwh atyourRISCorCISCpersuasion[1].thei960.Also. butanewtechnologycalledEPIC.and manyotherlow-costRISCprocessorsarenot goingtoeasilygainasignificantperformancead vantageoverfutureimplementationofthex86 architecture.evenifthebattlelinesarenow becomingfuzzy.duetothesiz elimitationofchips. CISCchipssuchasMotorola's68Kfamily providebettersoftwareavailabilityforsuchsyste m. ItseemslikelythatthePowerPC. Butthe futuremightnotbringvictorytooneofthem.itisunlikelytohavelarge memory.RISCgenerallydon'thaveadva ntages overCISC. Finally.sohighcodedensityisimportant.Inanembeddedsystem.Thisseemsclear.So.When chooseaprocessorforembeddedrealtimeapplication.sincemostofrealtimesystemsrequireveryfastinterrupthandling andhighcode density.

Becauseofthex8 6marketitisnotlikelythatCISCwilldiesoon.forExplicitlyParallelInstructionComputing. LikethewordparallelalreadysaysEPICcando manyinstructionexecutionsinparalleltoone another.whiletheRISC processorsarebecomingextinct. IfIntel'sEPICarchitectureissuccessful. Andfinally.Thiswillin theoryallowtheprocessingofWindowsbasedaswellasUNIXbasedapplicationsbythesame CPU.2Summary Fromourlimitedexperiencebasedontheresults ofourbenchmarks. 4.Microsoftisalreadydevelopingthe irWin64 standardforit. butRISCmay.itappearsthattheoretically .AllofthebigC PUmanufacturesbutSunandMotorolaarenow sellingx86basedproducts. EPICmightmakefirstRISCobsoleteandlater CISCtoo. IntelisworkingonitundercodenameMerced.Mercedwill bea64-bitchip.SGI). SothefuturemightbringEPICprocessorsand moreCISCprocessors. EPICisacreatedbyIntelandisinawayac ombinationofbothCISCandRISC.itmight bethebiggestthreadforRISC.andsomearejustwaitingforMe rcedtocomeout(HP.Likethenamesays.

Millerforhisexcellentteaching.however. Acknowledgements Wewouldliketoexpressoursinceregratitudet oDr.thebottomline betweenCISCandRISCbecomesblur.EthanL.inreal world.Wethinktheadoptionofeach other’stechnologytoovercomeitsown drawbacksmaybemoreandmoreatrendinfut ureCPUdesign.from . Withtimefleeting. timelyhelpandsupport.howfastachipcanexecuteth einstructionsitisgivenandhowwellitruns existingsoftware.peopleonlycareabouthowwellasyste m canservethem.andenlighteningdiscuss ion. References [1]DanielMann."Whythex86CISCbeatRISC" .Also.thepureRISCmachinesuchasMIPSR2000is amorepromisingstyleofcomputerdesign comparedtoIntel80386CISCChipatthatera.otherarchitectures differentfromCISCandRISCmayappear.

com/cp/html/choosingproc.zytec.D.http://www.25-33.zdnet.1980).D. Pages8-21 [7]JeffProsise.1985).Patterson.001.fu-berlin.htm [2]KeithDiefendorff.1994). Pages28-33 [3]Radin.Patterson.Sequin."The801Minicomputer".Commun.com/pcmag/pctech/content/14/18/tu 1418.html [8]MargaritaEsponda.pp.andDitzel.html [6]DavidA.39-47 [4]Patterson.Commun.15. ComputerArchitectureNews8:6(Oct."HistoryofthePowerPCar chitecture".“TheRISC Concept-ASurveyofImplementations”.de/lehre/WS94/RA/RISC9.“RISCvs.25 yearsoftheinternationalsymposiaonComputer .S.com/Benchmarks/whyx86.inf.CarloH.ACM28. pp."Thecasefo rthereducedinstructionsetcomputer".G.ACM37.“RISCI :AReducedInstructionSetVLSIComputer”.1(Jan.R.In ProceedingsoftheSymposiumonArchitectural SupportforProgrammingLanguagesandOperati ngSystems(March1982)."ChoosingaProcessorforEmb eddedReal-TimeApplications".from http://www.html [9]DavidA.Ra'ulRojas.from http://www. [5]DennisTerry.CISC:TheRealStory --WhatmakesthePowerPCaRISCprocessor andthePentiumaCISC?“.6(Jun.“ReducedInstructionSet Computers”.from http://www.amdembedded.

June27July2.Gross.JohnL.BarcelonaSpain.Volume6.ChristopherRowen.Pages229-257 .Issue3 (1988).Hennessy.Przybylski. “MeasurementandevaluationoftheMIPSarchite ctureandprocessor”.1998.architecture(selectedpapers).Steve nA.Pages216-230 [10]ThomasR.