电子信息工程专业专业英语三千字翻译.docx

上传人:b****2 文档编号:24538149 上传时间:2023-05-28 格式:DOCX 页数:9 大小:21.78KB
下载 相关 举报
电子信息工程专业专业英语三千字翻译.docx_第1页
第1页 / 共9页
电子信息工程专业专业英语三千字翻译.docx_第2页
第2页 / 共9页
电子信息工程专业专业英语三千字翻译.docx_第3页
第3页 / 共9页
电子信息工程专业专业英语三千字翻译.docx_第4页
第4页 / 共9页
电子信息工程专业专业英语三千字翻译.docx_第5页
第5页 / 共9页
点击查看更多>>
下载资源
资源描述

电子信息工程专业专业英语三千字翻译.docx

《电子信息工程专业专业英语三千字翻译.docx》由会员分享,可在线阅读,更多相关《电子信息工程专业专业英语三千字翻译.docx(9页珍藏版)》请在冰豆网上搜索。

电子信息工程专业专业英语三千字翻译.docx

电子信息工程专业专业英语三千字翻译

Unit3computerarchitectureandmicroprocessors

3--1ComputerArchitecture

1)Computerarchitecture,incomputerscience,isageneraltermreferringtothestructureofallorpartofcomputersystem.Thetermalsocoversthedesignofsystemsoftware,suchastheoperatingsystem(theprogramthatcontrolsthecomputer),aswellasreferringtothecombinationofhardwareandbasicsoftwarethatlinksthemachinesonacomputernetwork.Computerarchitecturereferstoanentirestructureandtothedetailsneededtomakeitfunctional.Thus,computerarchitecturecoverscomputersystems,microprocessors,circuits,andsystemprograms.Typicallythetermdoesnotrefertoapplicationprograms,suchasspreadsheetsorwordprocessing,whicharerequiredtoperformataskbutnottomakethesystemrun.

2)1.DesignElements

3)Indesigningacomputersystem,architectsconsiderfivemajorelementsthatmakeupthesystem'shardware:

thearithmetic/logicunit,controlunit,memory,input,andoutput.Thearithmetic/logicunitperformsarithmeticandcomparesnumericalvalues.Thecontrolunitdirectstheoperationofthecomputerbytakingtheuserinstructionsandtransformingthemintoelectricalsignalsthatthecomputer'scircuitrycanunderstand.Thecombinationofthearithmetic/logicunitandthecontrolunitiscalledthecentralprocessingunit(CPU).Thememorystoresinstructionsanddata.Theinputinputandoutputsectionsallowthecomputertoreceiveandsenddata,respectively.

4)Differenthardwarearchitecturesarerequiredbecauseofthespecializedneedsofsystemsandusers.Oneusermayneedasystemtodisplaygraphicsextremelyfast,whileanothersystemmayhavetobeoptimizedforsearchingadatabaseorconservingbatterypowerinalaptopcomputer.

5)Inadditiontothehardwaredesign,thearchitectsmustconsiderwhatsoftwareprogramswilloperatethesystem.Software,suchasprogramminglanguagesandoperatingsystems,makesthedetailsofthehardwarearchitectureinvisibletotheuser.Forexample,computersthatusetheCprogramminglanguageoraUNIToperatingsystemmayappearthesamefromtheuser'sviewpoint,althoughtheyusedifferenthardwarearchitectures.

6)2.ProcessingArchitecture

7)Whenacomputercarriesoutaninstruction,itproceedsthroughfivesteps.First,thecontrolunitretrievestheinstructionfrommemory,forexample,aninstructiontoaddtwonumbers.Second,thecontrolunitdecodestheinstructionintoelectronicsignalsthatcontrolthecomputer.Third,thecontrolunitfetchesthedata(thetwonumbers).Fourth,thearithmetic/logicunitperformsthespecificoperation(theadditionofthetwonumbers).Fifth,thecontrolunitsavestheresult(thesumofthetwonumbers).

8)Earlycomputersusedonlysimpleinstructionsbecausethecostofelectroniccapableofcarryingoutcomplexinstructionswashigh.Asthiscostdecreasedinthe1960s,morecomplicatedinstructionsbecamepossible.Complexinstructions(singleinstructionsthatspecifymultipleoperations)cansavetimebecausetheymakeitunnecessaryforthecomputertoretrieveadditionalinstructions.Forexample,ifsevenoperationsarecombinedinoneinstruction,thensixofthestepsthatfetchinstructionsareeliminatedandthecomputerspendslesstimeprocessingthatoperation.Computersthatcombineseveralinstructionsintoasingleoperationarecalledcomplexinstructionsetcomputers(CISC).

9)However,mostprogramsdonotoftenusecomplexinstructions,butconsistmostlyofsimpleinstructions.WhenthesesimpleinstructionsarerunonCISCarchitectures,theyslowdownprocessingbecauseeachinstruction--whethersimpleorcomplex--takeslongertodecodeinaCISCdesign.Analternativestrategyistoreturntodesignsthatuseonlysimple,single--operationinstructionsetsandmakethemostfrequentlyusedoperationsfasterinordertoincreaseoverallperformance.Computersthatfollowthisdesignarecalledreducedinstructionsetcomputers(RISC).

10)RISCdesignsareespeciallyfastatthenumericalcomputationsrequiredinscience,graphics,andengineeringapplications.CISCdesignsarecommonlyusedfornon-numericalcomputationsbecausetheyprovidespecialinstructionsetsforhandlingcharacterdata,suchastextinawordprocessingprogram.SpecializedCISCarchitectures,calleddigitalsignalprocessors,existtoaccelerateprocessingofdigitizedaudioandvideosignals.

11)3.OpenandClosedArchitectures

12)TheCPUofacomputerisconnectedtomemoryandtotheoutsideworldbymeansofeitheranopenoraclosedarchitecture.Anopenarchitecturecanbeexpandedafterthesystemhasbeenbuilt,usuallybyaddingextracircuitry,suchasanewmicroprocessorcomputerchipconnectedtothemainsystem.Thespecificationsofthecircuitryaremadepublic,allowingothercompaniestomanufacturetheseexpansionproducts.

13)Closedarchitecturesareusuallyemployedinspecializedcomputersthatwillnotrequireexpansion,forexample,computersthatcontrolmicrowaveovens.Somecomputermanufacturershaveusedclosedarchitecturessothattheircustomerscanpurchaseexpansioncircuitryonlyfromthem.Thisallowsthemanufacturetochargemoreandreducestheoptionsfortheconsumer.

14)4.NetworkArchitecture

15)Computerscommunicatewithothercomputersvianetworks.Thesimplestnetworkisadirectconnectionbetweentwocomputers.However,computerscanalsobeconnectedoverlargenetworks,allowinguserstoexchangedata,communicateviaelectronicmail,andshareresourcessuchasprinters.

16)Computerscanbeconnectedinseveralways.Inaringconfiguration,dataaretransmittedalongtheringandeachcomputerintheringexaminesthisdatatodetermineifitistheintendedrecipient.Ifthedataarenotintendedforaparticularcomputer,thecomputerpassesthedatatothenextcomputerinthering.Thisprocessisrepeateduntilthedataarriveattheirintendeddestination.Aringnetworkallowsmultiplemessagestobecarriedsimultaneously,butsinceeachmessageischeckedbyeachcomputer,datatransmissionisslowed.

17)Inabusconfiguration,computersareconnectedthroughasinglesetofwires,calledabus.Onecomputersendsdatatoanotherbybroadcastingtheaddressofthereceiveandthedataoverthebus.Allthecomputersinthenetworklookattheaddresssimultaneously,andtheintendedrecipientacceptsthedata.Abusnetwork,unlikearingnetwork,allowsdatatobesentdirectlyfromonecomputertoanother.However,onlyonecomputeratatimecantransmitdata.Theothersmustwaittosendtheirmessages.

18)Inastarconfiguration,computersarelinkedtoacentralcomputercalledahub.Acomputersendstheaddressofthereceiverandthedatatothehub,whichthenlinksthesendingandreceivingcomputersdirectly.Astarnetworkallowsmultiplemessagestobesentsimultaneously,butitismorecostlybecauseitusesanadditionalcomputer,thehub,todirectthedata.

19)5.RecentAdvances

20)OneproblemincomputerarchitectureiscausedbythedifferencebetweenthespeedoftheCPUandthespeedatwhichmemorysuppliesinstructionsanddata.ModernCPUscanprocessinstructionsin3nanoseconds(3billionthsofasecond).Atypicalmemoryaccess,however,takes100nanosecondsandeachinstructionmayrequiremultipleaccesses.Tocompensateforthisdisparity,newcomputerchipshavebeendesignedthatcontainsmallmemories,calledcaches,locatedneartheCPU.BecauseoftheirproximitytotheCPUandtheirsmallsize,cachescansupplyinstructionsanddatafasterthannormalmemory.Cachememorystoresthemostfrequentlyusedinstructionsanddataandcangreatlyincreaseefficiency.

21)Althoughalargecachememorycanholdmoredata,italsobecomesslower.Tocompensate,computerarchitectsemploydesignswithmultiplecaches.ThedesignplacesthesmallestandfastestcachenearesttheCPUandlocatesasecondlargeandslowercachefartheraway.ThisarrangementallowstheCPUtooperateonthemostfrequentlyaccessedinstructionsanddataattopspeedandtoslowdownonlyslightlywhenaccessingthesecondarycache.UsingseparatecachesforinstructionsanddataalsoallowstheCPUtoretrieveaninstructionanddatasimultaneously.

22)Antherstrategytoincreasespeedandefficiencyistheuseofmultiplearithmetic/logicunitsforsimultaneousoperations,calledsuperscalarexecution.Inthisdesign,instructionsareacquiredingroups.Thecontrolunitexamineseachgrouptoseeifitcontainsinstructionsthatcanbeperformedtogether.Somedesignsexecuteasmanyassixoperationssimultaneously.Itisrare,however,tohavethismanyinstructionsruntogether,soonaveragetheCPUdoesnotachieveasix-foldincreaseinperformance.

23)Multiplecomputersaresometimescombinedintosinglesystemscalledparallelprocessors.Whenamachinehasmorethanonethousandarithmetic/logicunits,itissaidtobemassivelyparallel.Suchmachinesareusedprimarilyfornumericallyintensivescientificandengineeringcomputation.Parallelmachinescontainingasmanyassixteenthousandcomputershavebeenconstructed.

3-3VLIWMicroprocessors

24)WhenTransmetaCorp.revealeditsnewCrusoeofprocessorslastmonth,expertsweren’tsurprisedtolearnthatthechipsarebasedonVeryLongInstructionWord(VLIW)technology.VLIWhasbecometheprevailingphilosophyofmicroprocessordesign,eclipsingolderapproachessuchasRISCandcomplexinstructionsetcomputing(CISC).

25)Allmicroprocessordesignsseekbetterperformancewithinthelimitationsoft

展开阅读全文
相关资源
猜你喜欢
相关搜索

当前位置:首页 > 自然科学 > 化学

copyright@ 2008-2022 冰豆网网站版权所有

经营许可证编号:鄂ICP备2022015515号-1