《微机原理及应用实践(英文版)》Chapter 3 　INSTRUCTION SYSTEM

Chapter3INSTRUCTIONSYSTEM3.1AddressingModes3.2InstructionFormandCode3.3InstructionSet3.4TroubleShootingTechniquesOBJECTIVES

Uponfinishingthischapter,youwilllearn:

Thewaysthatmicroprocessorslocatethedatawheretobestoredandhowtobebroughtbyaninstruction.

Instructioncodeshowtoencodeanddecodeby80x86.

3.1AddressingModes

Amicroprocessorinstructionisabinaryword.Whenitisreadasaninstruction,thisbinarywordtellsthemicroprocessortodoonesimpletask.Nootherbinarywordtellsthemicroprocessortodothistask.Mostmicroprocessorinstructionsletyoumoveorprocessdata.Thedatamaybeinmemory,oritmaybeinoneofthemicroprocessor’sregisters.Afewothermicroprocessorinstructionsletyoudo“housekeeping”functions.

Theseinstructionsletyoucontrolcertainmicroprocessorinstructions.1Aninstructioncanbebrokenintotwoparts,calledtheopcode(operationcode)andtheoperands.Theopcodetellsthemicroprocessorwhattodo,andtheoperandtellsitthecontentofthisdoingaction.Oftenyouwillfindthatthefirstwordofamultiwordinstructionistheopcode.Eitherthesecondwordorthesecondandthirdwordsareoperands.

Themicroprocessor’saddressingmodesarethewaythatthemicroprocessoridentifiesthelocationoftheoperandsordata.Whenthe8086executesaninstruction,itperformsthespecifiedfunctionindata.Thesedata,calledoperands,maybeonepartoftheinstruction,mayresideinoneoftheinternalregistersofthemicroprocessor,maybestoredatanaddressinmemory,ormaybeheldatanI/Oport.Aswewillsee,thenormalmicroprocessorhasmanydifferentwaysanditcanidentifydatainaparticularlocation.

Eachdifferentwaymaybeusedatadifferenttime,dependingonothercircumstances.2Toaccessthesedifferenttypesofoperands,the8086isprovidedwithvariousaddressingmodeswhicharecategorizedintothreetypes:registeroperandaddressing,immediateoperandaddressing,andmemoryoperandaddressing.Letusnowconsiderindetailtheaddressingmodesineachofthesecategories.3.1.1RegisterAddressing

Withtheregisteraddressingmode,theoperandtobeaccessedisspecifiedasresidinginaninternalregisterofthe8086.Table3.1liststheinternalregistersthatcanbeusedasasourceoradestinationoperand.Notethatonlythedataregisterscanbeaccessedaseitherabyteoraword.

Anexampleofaninstructionthatusesthisaddressingmodeis

MOVAX,BX

Thisstandsfor“moveorcopythecontentsofBX,whichisthesourceoperand,toAX,whichisthedestinationoperand.”Boththesourceanddestinationoperandshavebeenspecifiedasthewordcontentsofinternalregisterofthe8086.

Letusnowlookattheeffectofexecutingmoveinstructionintheregisteraddressingmove.InFigure3.1(a),weseethestateofthe8086justpriortofetchingtheinstruction.NotethatthelogicaladdressformedfromCSandIP(CS:IP)pointstothe“MOVAX,BX”instructionatphysicaladdress01000H.Thisinstructionisfetchedintothe8086’sinstructionqueue,whereitisheldwaitingtobeexecuted.

Priortoexecutionofthisinstruction,thecontentsofBXareABCDH,andthecontentsofAXrepresentadon’t-carestate.Theinstructionisreadfromtheoutputsideofthequeue,thendecoded,andexecuted.AsFigure3.1(b)shows,theresultproducedbyexecutingthisinstructionisthatthevalueABCDHiscopiedintoAX,andtheIP=IP+2pointsto01002H.Figure3.1(a)Registeraddressingmodeinstructionbeforefetchingandexecution；(b)Afterexecution

Single-byteinstructionsexecutefasterthananyotherinstructions.Themicroprocessormustcompletetheexecutionofaninstruction.Asingle-byteinstructionisexecutionofaninstruction，anditisexecutedintwomicroprocessorcycles.Thefetchingtakesonecycle,andtheexecutiontakestheothercycle.

Asyougoon,paycloseattentiontohowmanymicroprocessorcyclesareneededtocarryouttheinstructionsthatuseotheraddressingmodes.Youwillseethattheinstructionsusingtheinherentaddressingmodearethefastestinstructions.

3.1.2ImmediateAddressing

Immediateaddressingiseasytounderstand.Immediatelyfollowingtheopcodearethedata.Sotheinstructionwillbeencodedwithanimmediateoperand.Theremaybeeitheroneortwobytesofdata.Thesedataaresuppliedbytheprogrammer.Immediateaddressing,therefore,requiresnomemoryaddress.Theoperandisonepartoftheinstructioninsteadofthecontentsofaregisterormemorylocation,itrepresentswhatiscalledanimmediateoperand.

Sincethedataareencodeddirectlyintotheinstruction,immediateoperandsnormallyrepresentconstantdata.Immediateaddressingrequiresonlyanopcodeimmediatelyfollowedbythedatatobeoperatedon.Thisaddressingmodecanonlybeusedtospecifyasourceoperand.

Intheinstruction：

MOVAL,15H

Thesourceoperand15Hisanexampleofabyte-wideimmediatesourceoperand.Thedestinationoperand,whichisthecontentsofAL,usesregisteraddressing.Thus,thisinstructionemploysboththeimmediateandregistersaddressingmodes.

Figure3.2(a)and(b)illustratethefetchandexecutionofthisinstruction.Herewefindthattheimmediateoperand15Hisstoredinthecodesegmentofmemoryinthebytelocationimmediatelyfollowingtheopcodeoftheinstruction.Thisvalueisfetched,alongwiththeopcodeforMOV,intotheinstructionqueuewithinthe8086.Whenitperformsthemoveoperation,thesourceoperandisfetchedfromthequeue,notfromthememory,andnoexternalmemoryoperationsareperformed.Notethattheresultproducedbyexecutingthisinstructionisthattheimmediateoperand,whichequals15H,isloadedintothelower-bytepartoftheaccumulator(AL).Figure3.2(a)Immediateaddressingmodeinstructionbeforefetchandexecution；(b)Afterexecution

[ReadingMaterial]Whenisimmediateaddressinguseful?Agoodexampleisaprogramthatrequiresloadingtheaccumulatorwiththesame8-bitnumbereachtimetheprogramisrun.Intheinstruction’sfirstbyte,youputtheopcodethattellsthemicroprocessortoloaditsaccumulatorwiththecontentsofthebytethatimmediatelyfollowstheopcode.Inthatsecondbyte,youputthedatayouwantintotheaccumulator.

Theinstructionasisdescribedaboveiscalledtheregisterimmediateinstruction.Figure3.3showshowtheloadregisterimmediateinstructionappearswhenitisinmemory.Itshowsthe

opcodeintheinstruction’sfirstbyteandthedataintheinstruction’ssecondbyte.Ofcourse,boththeopcodeandthedatawillcontainbinaryinformationratherthanthenamesusedinFigure3.3.Figure.3.3Theloadregisterimmediateinstruction,wherethedataisinabytefollowingtheopcodebyte

Theimmediateaddressingmodeusestwomicroprocessorcycles.Duringthefirstcycle,theinstructionisfetched.Whentheinstructionisexecuted,thesecondbyte(thedata)isprocessedaccordingtowhatthefirstbyte(theopcode)tellsthemicroprocessor’slogictodo.

Inimmediateaddressing,thedatamustimmediatelyfollowtheinstructioninmemory.Thatis,thedataisactuallypartofthemicroprocessor’sinstruction.Butamicroprocessoroftenneedstouseinformationstoredatothermemorylocations.Insuchcasesimmediateaddressingwillnotworkinsuchcases,whichleadsustothenextformofaddressing.3.1.3MemoryAddressing

Toaccessanoperandinmemory,the8086/8088mustcalculatethephysicaladdressoftheoperandandtheninitiateareadorwriteoperationofthisstoragelocation.The8086/8088CPUisprovidedwithagroupofaddressingmodesknownasthememoryoperandaddressingmodesforthispurpose.LookingatFigure3.4,weseethatthephysicaladdressiscomputedfromasegmentbaseaddress(SBA)andaneffectiveaddress(EA).

SBAidentifiesthestartinglocationofthesegmentinmemory,andEArepresentstheoffsetoftheoperandfromthebeginningofthissegmentofmemory.EarlierweshowedhowSBAandEAarecombinedwithinthe8086/8088toformthelogicaladdressSBA:EAandhowtocomputethephysicaladdressfromthesetwovalues.

ThevalueoftheEAcanbespecifiedinavarietyofways.Onewayistoencodetheeffectiveaddressoftheoperanddirectlyintheinstruction.Thisrepresentsthesimplesttypeofmemoryaddressing,knownasthedirectaddressingmode.Figure3.4showsthataneffectiveaddresscanbemadeupfromasmanyasthreeelements:thebase,index,anddisplacement.Usingtheseelements,theeffectiveaddresscalculationismadebythegeneralformula

EA=Base+Index+Displacement

Figure3.4alsoidentifiestheregistersthatcanbeusedtoholdthevaluesofthesegmentbase,base,andindex.Forexample,3ittellsusthatanyofthefoursegmentregistercanbethesourceofthesegmentbaseforthephysicaladdresscalculationandthatthevalueofbasefortheeffectiveaddresscanbeineitherthebaseregister(BX)orbasepointregister(BP).Italsoidentifiesthesizespermittedforthedisplacement.Figure3.4Physicalandeffectiveaddresscomputationformemoryoperands

Notalltheseelementsarealwaysusedincalculatingtheeffectiveaddress.Infact,anumberofmemoryaddressingmodesaredefinedbyusingvariouscombinationsoftheseelementswhatwecanselectinFigure3.5.Theyarecalleddirectaddressing,registerindirectaddressing,basedaddressing,

indexedaddressing,

and

based-indexedaddressing.Forinstance,usingbasedaddressingmode,theeffectiveaddresscalculationincludesjustabase.Theseaddressingmodesprovidetheprogrammerwithdifferentwaysofcomputingtheeffectiveaddressofanoperandinmemory.Next,wewillexamineeachofthememoryoperandaddressingmodesindetail.

1.Directaddressing

Instructionsthatusedirectaddressingmayhave2or3bytes.Frequently,thesearereferredtoastwodifferentaddressingmodes.Thefirstbyteoftheseinstructionsistheopcode.Thesecondbyteand,ifpresent,thethirdbyteareamemoryaddress.Thememoryaddresspointstothememorylocationthatcontainsthedatatobeworkedon.Thesecondandthirdbytescanaddresstheentire65,536memorylocationsinan8-bitmicroprocessor.Ina16-bitmicroprocessor,thesecondandthirdwordsallowthemicroprocessortoaddressthefulladdressrangeof4Gbytes.

Directaddressingisinmanywaystheeasiestaddressingmodetounderstand.Ininherentaddressing,theaddressis“builtin”.Theprogrammerdoesnothavetoaddressamemorylocationtogetdata.Inimmediateaddressing,thedataisfoundinthebyteimmediatelyfollowingtheopcode.Again,theprogrammerdoesnothavetoaddressamemorylocationtogetthedata;thedataisbuiltintotheinstruction.Indirectaddressing,however,theprogrammeris,forthefirsttime,actuallygivingtheaddressofthememorylocationwheretheneededdataislocated.

AsFigure3.6shows,thiseffectiveaddressisuseddirectlyasthe16-bitoffsetofthestoragelocationoftheoperandfromthelocationspecifiedbythecurrentvalueintheselectedsegmentregisterinFigure3.5.ThedefaultsegmentisDS.Therefore,the20-bitphysicaladdressoftheoperandinmemoryisnormallyobtainedfromlogicaladdressDS:EA.But,byusingasegmentoverrideprefix(SOP)intheinstruction,anyofthefoursegmentregisterscanbereferenced.Figure3.5Specificationofadirectmemoryaddress

Anexampleofaninstructionthatusesdirectaddressingmodeforitssourceoperandis

MOVCX,[1234H]

Thisstandsfor“movethecontentsofthememorylocationwithoffset1234HinthecurrentdatasegmentintointernalregisterCX.”Theoffsetisencodedaspartoftheinstruction’smachinecode.

InFigure3.6(a),wefindthattheoffsetisstoredinthetwobytelocationsthatfollowtheinstruction’sopcode.Astheinstructionisexecuted,the8088combines1234Hwith0200Htogetthephysicaladdressofthesourceoperandasfollows:

PA=0200016+123416=0323416

Thenitreadsthewordofdatastartingatthisaddress,whichisBEEDH,andloadsitintotheCXregister.ThisresultisillustratedinFigure3.6(b).Figure3.6(a)Directaddressingmodeinstructionbeforefetchandexecution；(b)Afterexecution

Whenyouusedirectaddressing,youuseextramicroprocessorcycles.Inthemicroprocessorcycles,firstlythemicroprocessormustfetchtheinstruction’sopcode.Whentheinstructionisdecoded,themicroprocessorknowsthatitmustfetchanadditional2bytesofinformation.Thesebytesofdatamakeuptheaddressofthememorylocationwherethedataisstored.Eachadditionalfetchtakesonemoremicroprocessorcycle.

Oncetheopcodeandthe2addressbytesarefetched,themicroprocessorstillmustexecutetheinstruction.Thistakesafourthmicroprocessorcycle.

Fromthisyoucanseethatexecutinganinstructionwithdirectaddressingtakestwiceaslongasexecutinganinstructionwithimmediateaddressing.Ofcourse,directaddressingisnecessarywhenyouneedtoplacedataatanylocationinmemory.

[ReadingMaterial]Whenyouwriteprograms,youshoulduseasfewdirectaddressinginstructionsaspossiblebecausetheytakelongertoexecute.Thisisespeciallytruewhenyouarewritingspecificpartsofaprogramwhichmustoperateatveryhighspeeds.Frequentlywhenwritingthistypeofprogram,aprogrammerwillchangefromahigh-levellanguagetoanassembly-levellanguagetoensurethatuseoftime-consuminginstructions,suchasthedirectinstructions,islimited.

Asingledatawordinan8-bitmicroprocessorisnotlongenoughtoaddressa64Kwordmemorylocation.Inan8-bitwordusedinthesecondbyteoftheinstructioncanonlyaddress256memorylocations.Toovercomethis,aspecialopcodetellsyouthattheinstructionisaspecialformofdirectaddressinstructionwhichusesasingleaddressbyte.

Usually,thesesingle-byteaddressesstartatthe“bottom”ofmemory.Thatis,theystartataddresslocation0000H.Sometimesthissingle-byterangeofmemoryiscalledmemorypage0.Forexample,an8-bitmicroprocessorusingasinglebyteofaddressinformationmightaddressthefirst256memorylocations.Thesearememorylocations0to255(00H-FFH).Ontheotherhand,the2-byteaddresswordaddressesthefirst65,536(0000HtoFFFFH).

Whatistheadvantageofdirectaddressinglimitedtoone-wordaddresses?Theansweristhatitsavesmicroprocessorcycles.Itsavestime.Thiscanbeavaluableaddressingmodeforboth8-and16-bitmicroprocessorsiftherearefewbytesofdatathatmustbeaddressedfrequently.Thesefewbytesofdatacanbestoredinthefirst256bytesofmemoryinan8-bitmicroprocessor.Ofcourse,ina16-bitmicroprocessorthesesingle-addresswordinstructionscanaddressthefirst65,536memorylocations,andarethereforeveryuseful.

Howdoyoutellthetwoaddressingmodesapartifthemicroprocessorhasbothtwo-andthree-worddirectaddressinginstructions?Themanufacturergiveseachformofdirectaddressingadifferentname.Infact,manymicroprocessormanufacturerstreatthesetwoformsofdirectaddressingastwocompletelydifferentaddressingmodes.Theopcodesfortheseformsofaddressingarealsodifferent.Forexample,somemanufacturerscall2-bytedirectaddressingdirectaddressingofpage0addressingandrefertothe3-bytedirectaddressingasextendeddirectaddressing.

2. Registerindirectaddressing

Intheregisterindirectaddressingmode,theinstructioncontainstheopcodeandalsoindicateswhichregisterpointstothememorylocationcontainingtheneededdata.Thatis,theinstructiondoesnotcontainthedata,asinimmediateaddressing,ortheaddressofthedata,asindirectaddressing.Buttheinstructionindicateswhichregistercontainstheneededdata.Dependingonthesizeofthemicroprocessor’saddressspace,theregisterindirectaddressingmodemayindicatearegisterpairwhichpointstothecorrectmemoryaddress.

Registerindirectaddressingmodeissimilartothedirectaddressingwejustdescribedinbothofthem,whichmeansaneffectiveaddressiscombinedwiththecontentsofDStoobtainaphysicaladdress.However,theydifferinthewaythattheoffsetisspecified.Figure3.7showsthatthistimeEAresidesineitherabaseregisteroranindexregisterwithinthe8088.ThebaseregistercanbeeitherbaseregisterBXorbasepointerregisterBP,andtheindexregistercanbesourceindexregisterSIordestinationindexregisterDI.Useofasegmentoverrideprefixpermitsaccessofanothersegmentregister.Figure3.7Specificationofanindirectmemoryaddress

Anexampleis

MOVAX,[SI]

ExecutionofthisinstructionmovesthecontentsofthememorylocationthatisoffsetfromthebeginningofthecurrentdatasegmentbythevalueofEAinregisterSIintotheAXregister.

Forinstance,Figure3.8(a)and(b)showthatifSIcontains1234HandDScontains0200H,theresultafterexecutingtheinstructionisthatthecontentsofthememorylocationataddress

PA=0200016+123416=0323416

aremovedtotheAXregister.NoticeinFigure3.8(b)thatthisvalueisBEEDH.Inthisexample,thevalue1234HintheSIregistermusthavebeenloadedwithanotherinstructionpriortoexecutingthemoveinstruction.IfgiveMOVAX,ES:[SI],thenPA=120016+123416=1323416,AX=BAADHinsteadofBEEDH.Figure3.8(a)Instructionusingregisterindirectaddressingmodebeforefetchandexecution；(b)Afterexecution

Theresultproducedbyexecutingthisinstructionandthatfortheexampleforthedirectaddressingmodearethesame.However,theydifferinthewaytogeneratetheirphysicaladdress.ThedirectaddressingmethodlendsitselftoapplicationswherethevalueofEAisaconstant.Ontheotherhand,registerindirectaddressingcanbeusedwhenthevalueofEAiscalculatedandstored,forexample,inSIbyapreviousinstruction—thatis,EAisavariable.Forinstance,theinstructionsexecutedjustbeforeourexampleinstructioncouldhaveincrementedthevalueinSIbytwo.

[ReadingMaterial]Indirectaddressingisveryconvenientformovingdatatoandfromfrequentlyusedmemorylocations.Thisisespeciallytruewhenthedataisinlistorfileformat.Thatis,itisveryhelpfulwhenyouwanttostoreorreaddatainsequentialmemorylocations.

Aprogramthatfilesdatainsequentiallocations(onelocationafterthenext)isshowninFigure3.9.Thisprogramusesregisterindirectaddressing.ItinputsdatafromanI/Oporttotheaccumulator.Thedataisthenstoredinsequentialmemorylocations.Thisexamplemayuseinstructionsthatyoudonotyetfullyunderstand.Whatyouneedtoseehereishowtheindirectmemoryaddressingmodeisused.

Theprogramisshowninflowchartform.Intheflowcharts,eachnumberedboxshowstheprogramminglogicforoneinstruction.Followtheflowchartasyoureadthenumberedsteps.

Step①Thefirstinstructionestablishesthememorylocationwherethedatafilewillstart.Atthestartoftheprogram,youusealoadregisterimmediateinstructiontoloadtheregisterpairwiththefile’sstartingaddress.

Step②Wearenowreadytoloadtheaccumulator.ThesecondinstructioninputsdatafromtheI/Oportandplacesitintheaccumulator.TheinputinstructionbringsdatafromI/Oport01.

Step③Thisinstructionusesastoreinstructionwithanindirectaddress.Itstorestheaccumulator’sdatainthememorylocationpointedtobytheregisterpair.Note:Youdonotstorethedataintheregisterpair.Rather,thedataisstoredinthememorylocationpointedtobytheregisterpair.Youareusingregisterindirectaddressing.Oncethedataisstored,theprogramgetsanotherbyteofdataandstoresitinthenextmemorylocation.

Step④Tomaketheregisterpairpointtothenextmemorylocation,theregisterpairisincrementedbyonecount.

Step⑤Oncetheregisterpairisincremented,itpointstothememorylocationwherethenextbyteofdataistobestored.Youmaynowrepeattheinputandstoreinstructions.Ajumpinstructionreturnsyoutotheinputinstruction.Youjumpbacktotheinstructionthatfollowstheloadregisterpairimmediateinstruction.Youdothisbecauseyoudonotwanttoreloadtheregisterpair.Ifyouwentthroughtheloadprocess,youwoulddestroytheincrementedmemorypointerandkeeploadingdatainthesamememorylocation.Figure3.9Aprogramtoloaddataintosequentialmemorylocationsusingregisterindirectaddressing

3. Registerrelativeaddressing

LikeMOVAX,1234H[SI],relativeaddressingismuchlikeindexedaddressing,buttherearesomedifferences.First,theinstruction’soffsetisaddedtoadifferentregisterineachoftheseaddressingmodes.Inrelativeaddressing,theoffsetisrelativetotheprogramcounter’scurrentvalue.Second,relativeaddressingusuallyusesaspecialformofradixcomplementarithmetic.Theuseofthisspecialarithmeticmakespossiblebranchingforwardwhentheoffsethasalogic0inthemostsignificantbit.Thespecialarithmeticalsopermitsbranchingbackwardwhentheoffsethasalogic1inthemostsignificantbit.Theoffsetisthentreatedasaradixcomplementnumber.

[ReadingMaterial]Whydowesometimesuserelativeaddressinginsteadofindexedaddressing?Forsomemicroprocessors,relativeaddressingistheonlymodeprovideforsomeinstructions.Often,theconditionalbranchinstructionusesonlyrelativeaddressing.Furthermore,relativeaddressingpermitswritingposition-independentcode.Position-independentcodeisalsocalledrelocatableorrelativecode.

Theaddressingofprogrammingcodeiswrittenrelativetothedatainthemicroprocessor’sprogramcounter.Consequently,thesamecodecanbeusedatanymemorylocation.Figure3.10showsaflowchartforaroutinethatmaybecodedineitherabsoluteorrelativecode.Absolutecodemeansthatthecodemustbeusedatafixedsetofmemorylocations.Table3.2(a)showsabsolutecode,theTable3.2(b)showsrelativecode.Figure3.10Aflowchartofaprogramtotestaninputport’sstatusregister toseeifthedataavailableisset

Thissubroutineinputsdatafromastatusregister.TheANDimmediateinstructionmasksallbutthedataavailableflaginthestatusregister(statusregisterbit1).Ifthebitissettologic1,theaccumulatorwillnotbe0aftertheANDinstructionexecutes.Thismeansnojumptakesplace.ThenextinstructioninputsadatawordfromI/Oport01H.

Ifthestatusregisterbitisnotset,alltheaccumulatorbitsarelogic0.Ajumpoccurs.Hereiswherethetwoprogramsdiffer.TheprograminTable3.2(a)jumpsbacktomemorylocation0000H.However,theprograminTable3.2(b)usesrelativeaddressing.Itjumpsbacksixmemorylocationsformtheprogramcounter’scurrentvalue.Remember,theprogramcounterispointingtothein(next)instruction,whilethejumpinstructionisexecuting.

Thejumpinstructiontellsthemicroprocessortojumpbackbyasmanystepsastheradixcomplementofthenumberintheinstruction’ssecondbyte.Inthiscase,thebackwardjumpissixlocationsbecauseFAHistheradixcomplementof6.TheinstructionJZRFAreallyjustsays,“Subtract6fromthecontentsoftheprogramcounter.”Hereisthesubtraction:

Table3.3showstheprogramfromTable3.2(b).Here,theprogramstartsatmemorylocation01AEH.Noneoftheinstructionmnemonicsinthelistingnoranyoftheaddressesaredifferent.Jumpingbacksixlocationsrelativetotheprogramcounter,eventhoughtheprogramcounternowpointstomemorylocation01B4H,stilltakesyoubacktotheinputstatusregisterinstruction(nowatmemorylocation01AEH).

Asyouhaveseen,relativeaddressingfreestheprogrammerfromworryingaboutwheretheprogram’sabsoluteaddressis.Buttheprogrammermustbegoodathexadecimalarithmetictocomputetherelativeaddresses.Usinganassemblerrelievestheprogrammeroftheneedtocalculaterelativeaddresses.Theassemblerdoesallthecalculationsitself.

4. Basedindexedaddressing

Combiningthebasedaddressingmodeandtheindexedaddressingmoderesultsinanew,morepowerfulmodeknownasbased-indexedaddressingmode.Letusconsideranexampleofamoveinstructionusingthistypeofaddressing.

MOVAH,[BX][SI]

Notethatthesourceoperandisaccessedusingbased-indexedaddressingmode.Therefore,theeffectiveaddressofthesourceoperandisobtainedas

EA=(BX)+(SI)

AndthephysicaladdressoftheoperandiscomputedfromthecurrentsofDSandthecalculatedEA.

PA=DS:(BX)+(SI)

Figure3.11(a)and(b)presentanexampleofexecutingthisinstruction.Usingthecontentsofthevariousregistersintheexample,theaddressofthesourceoperandiscalculatedas

PA=0200016+100016+200016=0500016

ExecutionoftheinstructioncausesthevaluestoredatthislocationinmemorytobereadintoAH.Figure3.11(a) Instructionusingbased-indexedaddressingmodebeforefetchandexecution;(b) Afterexecution

4Thisaddressingmodecanbeusedtoaccesscomplexdatastructuressuchastwo-dimensionalarrays.Figure3.12(a)showshowitcanbeusedtoaccesselementsinanm×narrayofdata.Noticethatthedisplacement,whichisafixedvalue,locatesthearrayinmemory.Thebaseregisterspecifiesthemcoordinateofthearray,andtheindexregisteridentifiesthencoordinate.Simplychangingthevaluesinthebaseandindexregisterspermitsaccesstoanyelementinthearray.Figure3.12(b)showstheregisterpermittedinthebased-indexedphysicaladdresscomputation.Figure3.12(a)Based-indexedaddressingofatwo-dimensionalarrayofdata；

(b)Specificationofabased-indexedaddress

5. Relativebasedindexedaddressing

Thisaddressingmodecombinesallofthefeaturesoftheaddressingmodeswehavebeenexamining.Aswithalltheothers,itstillcanaccessmemorylocationsonlywithinone64KBsegment.Itdoes,however,givetheprogrammertheoptionofusing