Technical Reference Manual - Analog, Embedded Processing ...

1 TMS320C28xExtendedInstructionSetsTechnic alReferenceManualLiteratureNumber:SPRUHS 1 AMarch2014 (FPU).. the the C28xplus Viterbi,ComplexMathand CRCUnit-II(VCU-II).. the 32-BitAccessesto (VSTATUS).. (RB).. March2014 RevisedDecember2015 SubmitDocumentationFeedbackCopyright 2014 2015, (CRC) (TMU).. the C28x+ DelaySlot Operationson the March2014 RevisedDecember2015 ContentsSubmitDocumentationFeedbackCopyr ight 2014 2015, of StatusRegister(STF).. (RB).. + + FPU+ (VSTATUS).. (RB).. + FCU+ of Status(STF) (RB) (VSTATUS) (RB) DelaySlot(s).. ,source1, :Additionwith RightShiftand :Additionwith Requirementsfor of FiguresSPRUHS1A March2014 RevisedDecember2015 SubmitDocumentationFeedbackCopyright 2014 2015,TexasInstrumentsIncorporatedPreface SPRUHS1A March2014 RevisedDecember2015 ReadThisFirstThisdocumentdescribesthe architecture,pipeline,and instructionsetsof the TMU,VCU-II,and TMS320C2000 digitalsignalprocessor(DSP)platformis part of the TMS320 documentusesthe followingconventions.

2 Hexadecimalnumbersare shownwith the suffixh or with a leading0x. For example,the followingnumberis 40 hexadecimal(decimal64): 40h or 0x40. Registersin this documentare shownas figuresand describedin tables. Eachregisterfigureshowsa rectangledividedinto fieldsthat representthe fieldsof the is labeledwith its bit name,its beginningand endingbit numbersabove,and legendexplainsthe notationusedfor the properties Reservedbits in a registerfiguredesignatea bit that is usedfor followingbooksdescribethe TMS320x28xand relatedsupporttoolsthat are availableon the TIwebsite:DataManualand Errata SPRS439 TMS320F28335,TMS320F28334,TMS320F28332,T MS320F28235,TMS320F28234,TMS320F28232 DigitalSignalControllers(DSCs)DataManual containsthe pinout,signaldescriptions,as well as electricaland timingspecificationsfor the F2833 TMS320F28335,F28334,F28332,TMS320F28235, F28234,F28232 DigitalSignalControllers(DSCs)SiliconErr atadescribesthe advisoriesand usagenotesfor 'sGuides SPRU430 TMS320C28xCPUand InstructionSet ReferenceGuidedescribesthe centralprocessingunit (CPU)and the assemblylanguageinstructionsof the TMS320C28xfixed-pointdigitalsignalproces sors(DSPs).

3 It also describesemulationfeaturesavailableon TMS320C28xFloatingPointUnitand InstructionSet ReferenceGuidedescribesthefloating-point unit and includesthe instructionsfor the SPRU566 TMS320x28xx,28xxxDSPP eripheralReferenceGuidedescribesthe peripheralreferenceguidesof the 28x digitalsignalprocessors(DSPs).SPRUFB0 TMS320x2833x,2823xSystemControland InterruptsReferenceGuidedescribesthevari ousinterruptsand systemcontrolfeaturesof the 2833xand 2823xdigitalsignalcontrollers(DSCs).SPRU 812 TMS320x2833x,2823xAnalog-to-DigitalConve rter(ADC)ReferenceGuidedescribeshow to configureand use the on-chipADCmodule,whichis a TMS320x2833x,2823xDSCE xternalInterface(XINTF)ReferenceGuidedes cribestheXINTF,whichis a nonmultiplexedasynchronousbus, as it is usedon the 2833xand March2014 RevisedDecember2015 ReadThis FirstSubmitDocumentationFeedbackCopyrigh t 2014 2015, TMS320x2833x,2823xBootROMR eferenceGuidedescribesthe purposeand featuresofthe bootloader(factory-programmedboot-loadin gsoftware)and providesexamplesof alsodescribesothercontentsof the deviceon-chipbootROMand identifieswhereall of the informationis locatedwithinthat TMS320x2833x,2823xMultichannelBufferedSe rialPort(McBSP)ReferenceGuidedescribesth e McBSPavailableon the 2833xand McBSPsallowdirectinterfacebetweena DSPand otherdevicesin a TMS320x2833x,2823xDirectMemoryAccess(DMA )

4 ModuleReferenceGuidedescribesthe DMAon the 2833xand TMS320x2833x,2823xEnhancedPulseWidthModu lator(ePWM)ModuleReferenceGuidedescribes the mainareasof the enhancedpulsewidthmodulatorthat includedigitalmotorcontrol,switchmodepow ersupplycontrol,UPS(uninterruptiblepower supplies),and TMS320x2833x,2823xHigh-ResolutionPulseWi dthModulator(HRPWM)ReferenceGuidedescrib esthe operationof the high-resolutionextensionto the pulsewidthmodulator(HRPWM).SPRUFG4 TMS320x2833x,2823xEnhancedCapture(eCAP)M oduleReferenceGuidedescribesthe includesthe moduledescriptionand TMS320x2833x,2823xEnhancedQuadratureEnco derPulse(eQEP)ModuleReferenceGuidedescri besthe eQEPmodule,whichis usedfor interfacingwith a linearor rotaryincrementalencoderto get position,direction,and speedinformationfroma rotatingmachineinhigh-performancemotiona nd includesthe TMS320x2833x,2823xEnhancedControllerArea Network(eCAN)ReferenceGuidedescribesthe eCANthat usesestablishedprotocolto communicateseriallywith TMS320x2833x,2823xSerialCommunicationsIn terface(SCI)ReferenceGuidedescribesthe SCI,whichis a two-wireasynchronousserialport,commonlyk nownas a modulessupportdigitalcommunicationsbetwe enthe CPUand otherasynchronousperipheralsthat use the standardnon-return-to-zero(NRZ) TMS320x2833x,2823xDSCS erialPeripheralInterface(SPI)ReferenceGu idedescribesthe SPI - a high-speedsynchronousserialinput/output( I/O)

5 Port - that allowsa serialbitstreamof programmedlength(oneto sixteenbits) to be shiftedinto and out of the deviceat TMS320x2833x,2823xInter-IntegratedCircui t(I2C)ModuleReferenceGuidedescribesthe featuresand operationof the inter-integratedcircuit(I2C) SPRU513 'sGuidedescribesthe assemblylanguagetools(assemblerand othertoolsusedto developassemblylanguagecode),assemblerdi rectives,macros,commonobjectfile format,and symbolicdebuggingdirectivesfor TMS320C28xOptimizingC/C++ 'sGuidedescribestheTMS320C28x C/C++ compileracceptsANSI standardC/C++sourcecodeandproducesTMS320 DSPassemblylanguagesourcecodefor the TMS320C28xInstructionSet SimulatorTechnicalOverviewdescribesthe simulator,availablewithinthe CodeComposerStudiofor TMS320C2000 IDE,that simulatesthe instructionset of the C28x TMS320C28 (API) FirstSPRUHS1A March2014 RevisedDecember2015 SubmitDocumentationFeedbackCopyright 2014 2015, ,C28x,TMS320C2000are trademarksof March2014 RevisedDecember2015 ReadThis FirstSubmitDocumentationFeedbackCopyrigh t 2014 2015, FirstSPRUHS1A March2014 RevisedDecember2015 SubmitDocumentationFeedbackCopyright 2014 2015,TexasInstrumentsIncorporatedChapter 1 SPRUHS1A March2014 RevisedDecember2015 FloatingPointUnit(FPU)TheTMS320C2000 DSPfamilyconsistsof fixed-pointandfloating-pointdigitalsigna lcontrollers(DSCs).

6 TMS320C2000 DigitalSignalControllerscombinecontrolpe ripheralintegrationand easeofuseof a microcontroller(MCU)withthe processingpowerandC efficiencyof TI s overviewof the architecturalstructureand componentsof the C28xplus floating-pointunit the March2014 RevisedDecember2015 FloatingPointUnit (FPU)SubmitDocumentationFeedbackCopyrigh t 2014 2015,TexasInstrumentsIncorporatedProgram address bus (22)Program data bus (32)Read address bus (32)Read data bus (32)Write data bus (32)Existingmemory,peripherals,interface sPIEW rite address bus (32)LVFLUFC28x+ C28xplus floating-point(C28x+FPU)processorextends the capabilitiesof the C28xfixed-pointCPUby addingregistersand instructionsto devicedrawsfromthe best featuresof digitalsignalprocessing;reducedinstructi onset computing(RISC);andmicrocontrollerarchit ectures,firmware,and tool DSCfeaturesincludea modifiedHarvardarchitectureand RISC featuresare single-cycleinstructionexecution,registe r-to-registeroperations,and modifiedHarvardarchitecture(usablein Von Neumannmode).

7 Themicrocontrollerfeaturesincludeeaseof use throughan intuitiveinstructionset, byte packingandunpacking,and bit modifiedHarvardarchitectureof the CPUenablesinstructionanddatafetchesto be performedin CPUcan readinstructionsand datawhileit writesdatasimultaneouslyto maintainthe single-cycleinstructionoperationacrossth e CPUdoesthisoversix documentthe followingnotationsare used: C28xrefersto the C28xfixed-pointCPU. C28xplus Floating-Pointand C28x+FPUbothreferto the C28xCPUwith enhancementsto C28xFixed-PointCPUNo changeshavebeenmadeto the C28xbaseset of instructions,pipeline,or memorybus ,programswrittenfor the C28xCPUare completelycompatiblewith the C28x+FPUand all ofthe featuresof the C28xdocumentedinTMS320C28xDSPCPUand InstructionSet ReferenceGuide(literaturenumberSPRU430) applyto the C28x+ showsbasicfunctionsof the FPUF unctionalBlockDiagram10 FloatingPointUnit (FPU)SPRUHS1A March2014 RevisedDecember2015 SubmitDocumentationFeedbackCopyright 2014 2015, the C28xplus , or later,with at leastservicerelease8.

8 The ,or later,is also requiredto compileris availablevia CodeComposerStudioupdateadvisoras a generatebothfixed-pointas well as buildfloating-pointcodeuse thecompilerswitches:-v28and - -float_support= the float_supportoptionis in the buildoptionsundercompiler-> float_supportflag, or with float_support= none,the compilerwill C28xfloating-pointmakesureall associatedlibrarieshavealso beenbuilt for standardrun-timesupport(RTS)libariesbuil t for floating-pointincludedwith the compilerhavefpu32in for the floating-pointunit. The "eh" versionhas exceptionhandlingfor C++ fixed-pointRTSlibrariesin afloating-pointprojectwill resultin the linkerissuingan errorfor improveperformanceof nativefloating-pointprojects,considerusi ngtheC28xFPUFastRTSL ibrary(SPRC664). This librarycontainshand-codedoptimizedmathro utinessuchas division,squareroot,atan2,sin and cos. This librarycan be linkedinto yourprojectbeforethe standardruntimesupportlibraryto give yourapplicationa an example,the standardRTSlibraryusesa polynomialexpansionto calculatethe sin FastRTSlibrary,however,usesa mathlook-uptablein thebootROMof the look-uptablemethodresultsin approximatelya 20 cyclesavingsoverthe the C28xplusFloating-PointCPUThe C28x+FPUcontains: A centralprocessingunit for generatingdataand program-memoryaddresses;decodingand executinginstructions;performingarithmet ic,logical,and shift operations;and controllingdatatransfersamongCPUregister s,datamemory,and programmemory A floating-pointunit for IEEE single-precisionfloatingpointoperations.

9 Emulationlogicfor monitoringand controllingvariouspartsand functionsof the deviceand for logicis identicalto that on the C28xfixed-pointCPU. Signalsfor interfacingwith memoryand peripherals,clockingand controllingthe CPUand theemulationlogic,showingthe statusof the CPUand the emulationlogic,and logicis identicalto the the C28x+FPUcentralprocessingunit are: Fixed-Pointinstructionsare pipelinefor fixed-pointinstructionsis identicaltothat on the CPUimplementsan 8-phasepipelinethat preventsa writetoand a readfromthe samelocationfromoccurringout of Figure1-5. alignmentis donethroughsoftwaretoallowthe userto improveperformanceby takingadvantageof requireddelayslots. system-controlregisters,mathregisters, system-controlregistersare accessedby specialinstructions. Arithmeticlogicunit (ALU).The 32-bitALUperforms2s-complementarithmetic and Booleanlogicoperations. Floatingpointunit (FPU).The 32-bitFPUperformsIEEE single-precisionfloating-pointoperations .

10 Addressregisterarithmeticunit (ARAU).The ARAU generatesdatamemoryaddressesandincrement sor decrementspointersin parallelwith ALUoperations. shifterperformsall left and rightshiftsof can shift datato the leftby up to 16 bits and to the rightby up to 16 bits. multiplierperforms32-bit 32-bit2s-complementmultiplicationwith a multiplicationcan be performedwith two signednumbers,two unsignednumbers,or onesignednumberand one March2014 RevisedDecember2015 FloatingPointUnit (FPU)SubmitDocumentationFeedbackCopyrigh t 2014 2015,TexasInstrumentsIncorporatedCompone ntsof the C28xplus emulationlogicis identicalto that on the logicincludesthe followingfeatures: Debug-and-testdirectmemoryaccess(DT-DMA) .A debughost can gain directaccessto the contentof registersand memoryby takingcontrolof the memoryinterfaceduringunusedcyclesof theinstructionpipeline. A counterfor performancebenchmarking. of the followingdebugeventscan causea breakin programexecution: A breakpointinitiatedby the ESTOP0or ESTOP1instruction.