manualshive.com logo in svg

IBM PowerPC 604, Руководство пользователя

IBM PowerPC 604 - мощный процессор, созданный для высокопроизводительных компьютеров. У нас вы можете бесплатно скачать руководство пользователя для этого продукта. Посетите {сайт} и найдите нужное руководство в формате PDF для вашего удобства.

Поделиться
Скачать
background image

Содержание PowerPC 604

Страница 1: ......

Страница 2: ...Bus Interface Unit Operation Exceptions Memory Management Instruction Timing Signal Descriptions System Interface Operation Performance Monitor PowerPC Instruction Set Listings Invalid Instruction Fo...

Страница 3: ...e and Bus Interface Unit Operation Exceptions Memory Management Instruction Timing Signal Descriptions System Interface Operation Performance Monitor PowerPC Instruction Set Listings B Invalid Instruc...

Страница 4: ...MPR604UMU 01 MPC604UM AD 11 94 PowerPC 604 RISC Microprocessor User s Manual IBM Microelectronics...

Страница 5: ...or Obligation for damages ol any kind arising out ol the application or use ol these materials Any warranty or other Obligations as to the products described herein shall be undertaken solelyby Iha ma...

Страница 6: ...PowerPC 604 Microprocessor Features 1 2 PowerPC 604 Microprocessor Hardware Implementation 1 7 Instruction Flow 1 8 Fetch Unit 1 8 Decode Dispatch Unit 1 9 Branch Processing Unit BPU 1 9 Completion Un...

Страница 7: ...5 Calculating Effective Addresses 1 27 Exception Model 1 28 Instruction Timing 1 33 Power Management Nap Mode 1 35 Performance Monitor 1 35 Chapter 2 PowerPC 604 Processor Programming Model The PowerP...

Страница 8: ...al Instructions 2 28 Integer Rotate and Shift Instructions 2 29 Floating Point Instructions 2 30 Floating Point Arithmetic Instructions 2 30 Floating Point Multiply Add Instructions 2 31 Floating Poin...

Страница 9: ...r 3 Cache and Bus Interface Unit Operation 3 1 Data Cache Organization 3 3 3 2 Instruction Cache Organization 3 4 3 3 MMUs Bus Interface Unit 3 5 3 4 Memory Coherency Actions 3 8 3 4 1 604 Initiated L...

Страница 10: ...5 2 4 5 2 1 4 5 2 2 4 5 3 4 5 4 4 5 5 4 5 6 4 5 7 4 5 8 4 5 9 4 5 10 4 5 11 4 5 12 Contents Chapter 4 Exceptions PowerPC 604 Microprocessor Exceptions 4 2 Exception Recognition and Priorities 4 5 Exce...

Страница 11: ...ress Translation Selection 5 14 Selection of Page Address Translation 5 16 Selection of Direct Store Interface Address Translation 5 16 MMU Exceptions Summary 5 16 MMU Instructions and Register Summar...

Страница 12: ...Memory Operations 6 15 Write Back Mode 6 15 Write Through Mode 6 16 Cache Inhibited Mode 6 16 Timing Considerations 6 17 General Instruction Flow 6 17 Instruction Fetch Timing 6 18 Cache Hit Timing Ex...

Страница 13: ...iptions 7 3 Address Bus Arbitration Signals 7 3 Bus Request BR Output 7 4 Bus Grant BG lnput 7 4 Address Bus Busy ABB 7 5 Address Bus Busy ABB Output 7 5 Address Bus Busy ABB Input 7 5 Address Transfe...

Страница 14: ...l Output 7 15 Address Transfer Termination Signals 7 16 Address Acknowledge AACK Input 7 16 Address Retry ARTRY 7 16 Address Retry ARTRY Output 7 16 Address Retry ARTRY lnput 7 17 Shared SHD 7 17 Shar...

Страница 15: ...System Clock SYSCLK lnput 7 30 Test Clock CLK_OUT Output 7 31 Analog VDD AVDD lnput 7 31 PLL Configuration PLL_CFGO PLL_CFG3 Input 7 31 Chapter 8 System Interface Operation PowerPC 604 Microprocessor...

Страница 16: ...on Protocol Details 8 42 Packet 0 8 43 Packet 1 8 44 1 0 Reply Operations 8 44 Direct Store Operation Timing 8 46 Optional Bus Configuration 8 48 Fast L2 Data Streaming Mode 8 48 Fast L2 Data Streamin...

Страница 17: ...ervention 9 10 Warnings 9 11 Nonthreshold Events 9 11 Appendix A PowerPC Instruction Set Listings A 1 Instructions Sorted by Mnemonic A 1 A 2 Instructions Sorted by Opcode A 10 A 3 Instructions Groupe...

Страница 18: ...WIM 001 3 14 Machine Status Save Restore Register 0 4 6 Machine Status Save Restore Register 1 4 6 Machine State Register MSR 4 7 MMU Conceptual Block Diagram 32 Bit Implementations 5 6 PowerPC 604 Mi...

Страница 19: ...sor Bus for a Single Beat Transfer 8 6 Address Bus Arbitration 8 10 Address Bus Arbitration Showing Bus Parking 8 11 Address Bus Transfer 8 12 Snooped Address Cycle with ARTRY 8 19 Data Bus Arbitratio...

Страница 20: ...tructions 2 26 Integer Compare Instructions 2 28 Integer Logical Instructions 2 28 Integer Rotate Instructions 2 29 Integer Shift Instructions 2 30 Floating Point Arithmetic Instructions 2 30 Floating...

Страница 21: ...onse to Bus Transactions 3 20 3 5 604 Bus Operations Initiated by Cache Control Instructions 3 23 3 6 Cache Actions 3 24 4 1 Exception Classifications 4 3 4 2 Exceptions and Conditions Overview 4 3 4...

Страница 22: ...9 4 PMC2 Events MMCRO 26 31 Select Encoding 9 5 MMCRO Bit Settings 9 7 Complete Instruction List Sorted by Mnemonic A 1 Complete Instruction List Sorted by Opcode A 10 Integer Arithmetic Instructions...

Страница 23: ...structions A 27 A 31 I Form A 28 A 32 B Form A 28 A 33 SC Form A 28 A 34 D Form A 28 A 35 DS Form A 30 A 36 X Form A 30 A 37 XL Form A 34 A 38 XFX Form A 35 A 39 XFL Form A 35 A 40 XS Form A 36 A 41 X...

Страница 24: ...User s Manual summarizes features of the 604 that are not defined by the architecture This document and The Programming Environments Manual distinguishes between the three levels or programming envir...

Страница 25: ...ch chapters may include informationfrom multiple levels ofthe architecture For example the discussion of the cache model uses information from both the VEA and theOEA The information in this book is s...

Страница 26: ...form Appendix B Invalid Instruction Forms describes how invalid instructions are treated by the 604 This manual also includes a glossary and an index In this document the terms PowerPC 604 Microproces...

Страница 27: ...en they are low and negated when they are high Signals that are not active low such as APO AP3 address bus parity signals and TTO TT4 transfer type signals are referred to as asserted when they are hi...

Страница 28: ...dress space register BAT Block address translation BIST Buitt in seH test BIU Bus interface unit BHT Branch history table BPU Branch processing unit BTAC Branch target address cache BUID Bus unit ID C...

Страница 29: ...oup L2 Secondary cache LR Unk register LAU Least recently used LSB Least slgnHlcant byte lsb Least signHlcant bit LSU Loadlstore unit MCIU Multiple cycle Integer unit MESI ModHied exclusive sharedlinv...

Страница 30: ...anslation SIA Sampled instruction address register SIMM Signed immediate value SLB Segment look aside buffer SPR Special purpose register SPRGn Registers available for general purposes SR Segment regi...

Страница 31: ...ddress Physical address Relocation Translation Storage locations Memory Storage the act of Access For a detailed discussion ol how the terms interrupt and exception are used in this document see the i...

Страница 32: ...briefly how those units interact The 604 is an implementation of the PowerPCTM family ofreduced instruction set computer RISC microprocessors The 604 implements the PowerPC ArchitectureTM as it is sp...

Страница 33: ...variable sized block translation The TLBs and the cache use least recently used LRU replacement algorithms The 604 has a 64 bitexternal data bus and a 32 bit address bus The 604 interface protocol al...

Страница 34: ...er Fetcher Instruction Queue 8 word INSTRUCTION UNIT 64Bit 128Bit Dispatch Unit I BHT I FPR File 128Bit IMMU ISR 11 IBAT I I1TLB I Array 128Bit 32Bit B II 64Bit DMMU Tags 16 Kbyte I Cache Store Queue...

Страница 35: ...ute in the MCIU take multiple cycles to execute Each SCIU has a two entry reservation station to minimize stalls The MCIU has a two entry reservation station and provides early exit three cycles for 1...

Страница 36: ...struction and data caches Harvard architecture 16 Kbyte four way set associative instruction and data caches LRU replacement algorithm 32 byte eight word cache block size Physically indexed physical t...

Страница 37: ...ions Multiprocessing support features include the following Hardware enforced four state cache coherency protocol MESI for data cache Bits are provided in the instructioncache to indicate only whether...

Страница 38: ...tribute to the improved efficiency in instruction execution and more clearly indicating the relationships between execution units and their associated register files Branch Corraction Fetch Unit Dispa...

Страница 39: ...instruction queue by accessing the on chip instruction cache Typically the fetch unit continues fetching sequentially as many as four instructions at a time The address ofthe next instruction to be fe...

Страница 40: ...ion unit A reorder buffer ROB entry is allocated for each instruction and dependency checking is done between the instructions in the dispatch queue The rename buffers are searched for the operands as...

Страница 41: ...ervation station execution can begin The completion unit does not transfer instruction results from the rename registers to the registers lintil any speculative branch conditions preceding it in the c...

Страница 42: ...are written to the floating point rename buffers and to the reservation stations and are made available to subsequent instructions Instructions are executed from each reservation station in dispatch...

Страница 43: ...data structure that defines the mapping between virtual page numbers and physical page numbers The page table size is a power of 2 and its starting address is a multiple of its size Address translati...

Страница 44: ...cycle the data cache provides double word access to the LSU The data cache tags are dual ported so the process of snooping does not affect other transactions on the systeminterface Ifa snoop hit occu...

Страница 45: ...ace Unit BIU The 604 provides a versatile bus interface that allows a wide variety of system design options The interface includes a 72 bit data bus 64 bits of data and 8 bits of parity a 36 bit addre...

Страница 46: ...n the line are fetched later The critical double word as well as other words in the cache block are forwarded to the fetcher or to the LSU before they are written to the cache Memory accesses canoccur...

Страница 47: ...ts in one bus clock cycle Data transfers occur in either single beat transactions or four beat burst transactions Asingle beat transaction transfers as much as 64 bits Single beat transactions are cau...

Страница 48: ...gnals These signals are used to interrupt and under various conditions to reset the processor Processor state signals These two signals are used to set the reservation coherency bit and set the size o...

Страница 49: ...STATUS 1PROCESSOR JCONFIGURATION c s __ JTAGICOP J INTERFACE Figure 1 5 PowerPC 604 Microprocessor Signal Groups 1 2 6 Clocking The 604 has a phase locked loop PLL that generates the internal processo...

Страница 50: ...e base user level instruction set user level registers data types memory conventions and the memory and programming models seen by application programmers Note that the PowerPC architecture refers to...

Страница 51: ...o register operations for most computational insttuctions Source operands for these insttuctions are accessed from the registers or are provided as immediate values embedded in the insttuction opcode...

Страница 52: ...T2U DBAT2L DBAT3U DBAT3L SPR536 SPR537 SPR538 SPR539 SPR540 SPR541 SPR542 SPR543 Performance Monitor Segment Registers SRO SRI i SR15 SDR1 SDR1 ISPR25 Monitor Sampled Data Counters1 Monitor Control1 I...

Страница 53: ...provide a way to buffer data intended for the GPRs reducing stalls when the results ofone instruction are required by a subsequent instruction The use of rename buffers is not defined by the PowerPC a...

Страница 54: ...status configuring the processor and performing special operations Some SPRs are accessed implicitly as part of executing certain instructions All SPRs can be accessed by using the move to from SPR i...

Страница 55: ...gister Theregisters are implemented as a64 bitcounter with the least significantbitbeing the most frequently incremented The PowerPC architecture defines that the time base frequency be provided as a...

Страница 56: ...types permitting efficient decoding to occur in parallel with operand accesses This fixed instruction length and consistent format greatly simplifies instruction pipelining 1 3 3 1 1 Instruction Set T...

Страница 57: ...ns trap instructions and other instructions that affect the instruction flow Branch and trap instructions System call and rfi instructions Condition register logical instructions Synchronization instr...

Страница 58: ...et of 32 GPRs It also provides for word and double word operand loads and stores between memory and a set of 32 FPRs Computational instructions do not modify memory To use a memory operand in a comput...

Страница 59: ...a particular PowerPC processor may recognize exception conditions out of order exceptions are handled strictly in order When an instruction caused exception is recognized any unexecuted instructions t...

Страница 60: ...and any exceptions associated with those instructions complete execution These exceptions are maskable by setting MSR EE Asynchronous nonrnaskable There are two nonrnaskable asynchronous exceptions t...

Страница 61: ...using Conditions Type hex Reserved 00000 System reset 00100 Asystem reset is caused by the assertion of either the soft reset or hard reset signal Machine check 00200 A machine check exception is sign...

Страница 62: ...ess cannot be translated That is there is a page fault for this portion of the translation so an ISi exception must be taken to retrieve the translation from a storage device such as a hard disk drive...

Страница 63: ...rap A trap type program exception is generated when any of the conditions specified in a trap instruction is met Floating point 00800 Afloating point unavailable exception Is caused by an attempt to e...

Страница 64: ...de intem1pt which is described in Section 1 4 Power Management Nap Mode Reserved 01500 02FFF Reserved implementation specHic exceptions These are not implemented in the604 1 3 5 Instruction Timing As...

Страница 65: ...the results into the appropriate rename bufferentry andnotifies the completionstage thatthe instruction has finished execution The execution unit reports any internal exceptions to the completion sta...

Страница 66: ...nal to be negated Nap mode is exited clocks resume and MSR POW cleared when any asynchronous interrupt is detected 1 5 Performance Monitor The 604 incorporates a performance monitor facility that syst...

Страница 67: ...2 PowerPC Register Set in The Programming Environments Manual Note that registers are defined at all three levels of the PowerPC architecture user instruction set architecture VISA virtual environment...

Страница 68: ...t of the execution of an instruction Some registers are accessed both explicitly and implicitly The number to the right of the special purpose registers SPRs indicates the number that is used in the s...

Страница 69: ...40 DBAT2L SPR541 DBAT3U SPR542 DBAT3L SPR543 Segment Registers SRO SR1 I SR15 SDR1 SDR1 lsPR25 Performance Monitor Performance Monitor Mode Control Sampled Data Monitor Counters1 Register 01 Instructi...

Страница 70: ...Environments Manual Implementation Note The PowerPC architecture indicates that in some implementations the Move to ConditionRegister Fields mtcrt instructionmay performmore slowly when only aportion...

Страница 71: ...Aintroduces the time base facility TB a 64 bit structure that maintains and operates an interval timer The TB consists oftwo 32 bitregisters time baseupper TBU and time base lower TBL Note that the ti...

Страница 72: ...and IBATOL IBAT3L and four pairs of data BATs DBATOU DBAT3U and DBATOL DBAT3L See Figure 2 1 for a list of the SPR numbers for the BAT registers For more information see BAT Registers in Chapter 2 Po...

Страница 73: ...time base registers can be accessed by bothuser and supervisor level instructions See T IDle BaseFacility TB OEA in Chapter 2 PowerPC Register Set of The Programming Environments Manual for more infor...

Страница 74: ...ter 0 MMCRO This is used for enabling various performance monitoring interrupt conditions and establishes the function of the cpunters Sampled instruction address and sampled data address registers SI...

Страница 75: ...instruction address breakpoint exception is executed before the exception handler is invoked For more information about the IABR exception see Section 4 5 14 Instruction Address Breakpoint Exception O...

Страница 76: ...or does not cause a machine check exception 1 Enables the entry Into a machine check exception based on the detection of an address parity error Note that the machine check exception is further affect...

Страница 77: ...time Accesses to the cache from the bus are signaled as a miss while the invafidate all operation is in progress The bit is deared when the invalidation operation begins usually the cyde immediately f...

Страница 78: ...rs can be changed by hardware 1 If MSR PM is cleared the PMCn counters are net changed by hardware 5 ENINT Enable performance monitoring interrupt signaling 0 Interrupt signaling is disabled 1 Interru...

Страница 79: ...m for counting after a certain condition occurs or after a preset time has elapsed It can be used to support getting the count associated with a specific event 19 25 PMC1SELECT PMC1 input selector 128...

Страница 80: ...eady for execution 000 1001 Number of load deache misses that exceeded the threshold value with lateral L2 intervention 000 1010 Number of store deache misses that exceeded the threshold value with la...

Страница 81: ...0000 Instructions dispatched to the branch unit 01 0001 Instructions dispatched to the SCIUO 01 0010 Number cl loads completed 01 0011 Instructions dispatched to the MCIU 01 0100 Number cl snoop hit o...

Страница 82: ...The SDA can be read by using the mfspr instruction and written to by using the mtspr instruction SPR 959 2 2 Operand Conventions This section describes the operand conventions as they are represented...

Страница 83: ...length is implicit for each instruction 2 2 3 Alignment and Misaligned Accesses The operand of a single register memory access instruction has a natural alignment boundary equal to the operand length...

Страница 84: ...malized Single denormalized Normalize B and C ZeroBandC Double denormalized Double denormalized Single denormalized Normalized or zero Single denormalized Normalize A and C Zero A and C Double denorma...

Страница 85: ...nt location and alignment of operands in memory may affect the relative performance ofmemory accesses The best performance is guaranteed if memory operands are aligned on natural boundaries To obtain...

Страница 86: ...ptional External Control Instructions Note that this grouping of instructions does not necessarily indicate the execution unit that processes a particular instruction or group ofinstructions This info...

Страница 87: ...bits in reserved fields the results on execution can be said to be boundedly undefined If a user level program executes the incorrectly codedinstruction the resulting undefined results are bounded in...

Страница 88: ...ns Sorted by Opcode and Section 2 3 1 4 Reserved Instruction Class Notice that extended opcodes for instructions defined only for 64 bit implementations are illegal in 32 bit implementations and vice...

Страница 89: ...lating effective addresses as defined by the PowerPC architecture for 32 bit implementations For more detailed information see Conventions in Chapter 4 Addressing Modes and Instruction Set Summary of...

Страница 90: ...it 0 is ignored Load and store operations have three categories of effective address generation Register indirect with immediate index mode Register indirect with index mode Register indirect mode Ref...

Страница 91: ...invoked Exceptions can be caused directly by the execution of an insuuction as follows An attempt to execute an illegal insuuction causes the illegal insuuction program exception handler to be invoke...

Страница 92: ...ser level cache control synchronization and time base instructions user level registers programming model data types and addressing modes This section discusses the instructions defined in the UISA 2...

Страница 93: ...vwuo dlvwuo rD rA rB Although there is no Subtract Immediate instruction its effect can be achieved by using an addi instruction with the immediate operand negated Simplified mnemonics are provided th...

Страница 94: ...ber For information on simplified mnemonics for the integer compare instructions see Appendix F Simplified Mnemonics in The Programming Environments Manual 2 3 4 1 3 Integer Logical Instructions The l...

Страница 95: ...egister The result ofthe rotation is either inserted into the target register under control ofa mask ifa mask bit is 1 the associated bit of the rotated data is placed into the target register and if...

Страница 96: ...ns See Section 2 3 4 3 Load and Store Instructions for information about floating point loads and stores The PowerPC architecture supports a floating point system as defined in the IEEE 754 standard b...

Страница 97: ...multiply add instructions are summarized in Table 2 15 Table 2 15 Floatlng Polnt Multlply Add Instructions Name Mnemonic Operand Syntax Floating Multiply Add Double Precision fmadd lmadd frD frA lrC f...

Страница 98: ...Re bit set can cause an illegal instruction program exception or produce a boundedly undefinedresult In the 604 crfD should be treated as undefined 2 3 4 2 5 Floating Point Status and Control Register...

Страница 99: ...g Integer load instructions Integer store instructions Integer load and store with byte reverse instructions Integer load and store multiple instructions Floating point load instructions Floating poin...

Страница 100: ...lue The floating point load and store indexed instructions lfsx lfsux lfdx lfdux stfsx stfsux stfdx stfdux are also invalid when the Re bit is one In the 604 executing one of these invalid instruction...

Страница 101: ...n Notes The following notes describe the 604 implementation of integer load instructions In the PowerPC architecture the Re bit must be zero for almost all load and store instructions Ifthe Re bit is...

Страница 102: ...r integer store instructions the contents of rS are stored into the byte half word word or double word in memory addressed by the EA effective address Many store instructions have an update form in wh...

Страница 103: ...d value For the store with update instructions stbu stbux sthu sthux stwu stwux stfsu stfsux stfdu stfdux when rA 0 the instruction form is considered invalid In this case the 604 sets GPRO to an unde...

Страница 104: ...oss a page boundary However a OSI exception may occur when the boundary is crossed for example if a protection violation occurs on the new page Executing an lmw instruction in which rA is in the range...

Страница 105: ...ons In other PowerPC implementations operating with little endian byte order execution of a load or string instruction causes the system alignment error handler to be invoked see Section 3 2 2 Byte Or...

Страница 106: ...operation is not word aligned The 604 executes store string operations to cacheable memory at arate of one cycle per word if they are word aligned Cacheable store string operations that are not word...

Страница 107: ...th Update lfdu frD d rA Load Floating Point Double with Update Indexed lfdux frD rA rB 2 3 4 3 9 Floating Point Store Instructions This section describes floating point store instructions There are th...

Страница 108: ...e 2 27 shows the conversions made by the LSU when performing a Store Floating Point Single instruction Table 2 27 Store Floatlng Polnt Slngle Behavior FPR Precision Data Type Action Single Normalized...

Страница 109: ...on the value to be stored Because of how floating point numbers are implemented in the 604 there is also a case when execution of a store floating point double stfd stfdu stfdx stfdux instruction can...

Страница 110: ...e of bits in the CR Whenever the CR bits resolve the branch direction is either marked as correct or mispredicted Correcting a mispredicted branch requires that the 604 flush speculatively executed in...

Страница 111: ...Register OR with Complement crorc crbD crbA crbB Move Condition Register Field mcrf crfD crfS Note that if the LR update option is enabled for any of these instructions the PowerPC architecture define...

Страница 112: ...g to the condition register Table 2 33 Move to from Condition Register Instructions Name Mnemonic Operand Syntax Move to Condition Register Fields mtcrf CRM rS Move to Condition Register from XER mcrx...

Страница 113: ...Purpose Register mtspr SPR rS Move from Special Purpose Register mfspr rD SPR 2 3 4 7 Memory Synchronization lnstructions UISA Memory synchronization instructions control the order in which memory ope...

Страница 114: ...ructions Ifthe Re bit is one the instruction form is invalid These include the sync and lwarx instructions In the 604 executing one ofthese invalid instructionforms causes CRO to be set to anundefined...

Страница 115: ...is enabled by assertion of the timebase enable TBE input signal 2 3 5 2 Memory Synchronization Instructions VEA Memory synchronization instructions control the order in which memory operations are com...

Страница 116: ...management instructions defined by the VEA See 2 3 6 3 Memory Control Instructions OEA for information about supervisor level cache segment register manipulation and translation lookaside buffer mana...

Страница 117: ...es of the block that may reside in their cache this Is the kiU operation on the bus After it has exclusive access the 604 writes all zeros Into the cache block If the 604 already has exclusive access...

Страница 118: ...rS rA rB The eciwx and ecowx instructions cause an aligmnent exception if they are not word aligned 2 3 6 PowerPC OEA Instructions The PowerPC operating environment architecture OEA includes the struc...

Страница 119: ...000 MMCRO 953 11101 11001 PMC1 954 11101 11010 PMC2 955 11101 11011 SIA 959 11101 11111 SDA 1010 11111 10010 IABR 1023 11111 11111 PIR 1Note that the order of the two 5 bit halves of the SPR number Is...

Страница 120: ...4 Cache Management Supervisor Level Instruction Name Mnemonic Operand Syntax Implementation Notes Data debI rA rB The EA is computed translated and checked for protection Cache violations as defined i...

Страница 121: ...ns Name Mnemonic Operand Syntax Move to Segment Register mtsr SR rS Move to Segment Register Indirect mtsrln rS rB Move from Segment Register mfsr rD SR Move from Segment Register Indirect mfsrln rD r...

Страница 122: ...essors on the bus accept the operation AFffFIV is not asserted Once accepted the TLB invalidation is performed unless the processor is executing a multiple string instruction in which case the TLB inv...

Страница 123: ...ding asetofalternative mnemonics is providedfor some frequently usedoperations suchas no op loadimmediate loadaddress moveregister and complementregister Programs written to be portable across the var...

Страница 124: ...oherency at the page and block level as defined by the PowerPC architecture The caches use a least recently used LRU replacement policy The 604 cache implementation has the following characteristics S...

Страница 125: ...the cache Writes of cache blocks by the 604 for a copy back operation always present the first address of the block and transfer data beginning at the start ofthe block However this does not preclude...

Страница 126: ...s PA 0 19 PA 0 31 Cache Tags MMU Bus Interface Unit BIU Figure 3 2 cache Integration 3 1 Data Cache Organization Load Store Unit LSU Data 0 63 Data Cache 16 Kbyte Four Way SetAssociative Data 0 63 As...

Страница 127: ...ysically indexed The organization of the instruction cache shown in Figure 3 1 is identical to that of the data cache Each cache block contains eight contiguous words from memory that are loaded from...

Страница 128: ...re they are written to the cache The bus can be run at lx 2 3x l 2x or l 3x the speed of the processor The programmable on chip phase locked loop PLL generates the necessary processor clocks from the...

Страница 129: ...For a line fill operation the line fill address from either the instruction or data cache is kept in the memory address queue until the address can be sent out in an address tenure After the address...

Страница 130: ...ata is kept in the write bufferuntil bothcan be sent out in a write transaction Similarly for copy back operations the address is kept in the copy back address queue and the data is kept in the copy b...

Страница 131: ...ovide an overview of the behavior of the 604 with respect to load and store operations Table 3 1 does not include noncacheable cases The first three cases load when the cache block is marked I also in...

Страница 132: ...resolved before memory accesses that miss in the cache are forwarded onto the memory queue for arbitration onto the bus In addition although subsequentmemory accesses can address the cache full cohere...

Страница 133: ...if multiple processors are performing these types of memory operations to the same addresses without properly synchronizing one another through the use of the lwarx stwcx instructions the results of...

Страница 134: ...n the cache and in at least one other cache This blockis always consistent with system memory That is the shared state Is shared unmodilled there is no shared modilled state Invalid I This state Indic...

Страница 135: ...th the snooping operation The instructionprocessing is affectedonly when the snoop conttol logic detects a situation where a snoop push of modified data is required to maintain memory coherency M E 3...

Страница 136: ...defines two of the possible eight decodings of these bits to be unsupported WIM 110 or 111 Note that software must exercise care with respect to the use of these bits if coherent memory support is des...

Страница 137: ...e following coherency paradoxes can be encountered within a single processor Load or store operations to a page with WIM ObOll and a cache hit occurs Caching was supposed to be inhibited for this page...

Страница 138: ...These are described as follows Bit 1 Enable cache parity checking Enables a machine check exception based on the detection of a cache parity error If this bit is cleared cache parity errors are ignor...

Страница 139: ...he write to the register Any accesses to e cache from the bus are signaled as a miss during the time that the invalidate all operation is in progress The HIDO register can be accessed with the mtspr a...

Страница 140: ...execution of the debt instruction affects the state of the TLB and cache LRU bits as defined by the LRU algorithm 3 8 4 Data Cache Block Set to Zero dcbz As defined in the VEA when the debz instructi...

Страница 141: ...9 Basic Cache Operations This section describes operations that can occur to the cache and how these operations are implemented in the 604 3 9 1 Cache Reloads A cache block is reloaded after a read m...

Страница 142: ...the write on the system interface or the lwarx stwcx instructions The 604 drives two snoop status signals ARTRY and SHD in response to a qualified snoop request that hits These signals provide infonn...

Страница 143: ...issued by a processor after executing a dcbl instruction a dcbz instruction to a location marked I or S or a write operation to a block marked S If a kill blocktransaction appears on the bus and the...

Страница 144: ...date operation Is detected For more information on the tlble Instruction see Section 2 3 6 3 3 Translation Lookaside Buffer Management lnstructions OEA VO reply The 110 reply operation is part al the...

Страница 145: ...tanding on the bus and another pipelined bus operation hits against a modified block the 604 provides a high priority push operation This transaction can be enveloped within the address and data tenur...

Страница 146: ...e 3 5 does not include noncacheable or write through cases nor does it completely describe the mechanisms for the operations described For more information see Section 3 10 Cache Actions Chapter 3 Add...

Страница 147: ...k S 001 I Load Read 001 01010 nla XRTRVor Release the bus 7 RTRV Hl retry the operation 001 ME Load None n a n a nla n a Load from cache s 011 ESI Load Single 01M 01010 nla None or Load from main memo...

Страница 148: ...ead 000 11010 Set ii Load the block of data into cache atomic by set reservation this op load from cache mark cache block S 000 I lwanc Read 000 11010 Va AlmNor Release the bus atomic Am RV m retry th...

Страница 149: ...from main memory 011 M lwanc Single 01M 11010 nla ARTF Vor Paradox cache should be I 010 beat read ARTFIV m release the bus atomic retry the operation 100 I lwanc nla nla nla nla nla Alwanc to a page...

Страница 150: ...None or Store to main memory 010 flush 11M SRn 110 111 011 I Store Write with 01M 00010 n a Am RVor Release the bus 010 flush 11M XR fRV m retry the operation 110 111 011 ES Store Write with 01M 00010...

Страница 151: ...AD retry the oP ration 000 s stwcx Kill 000 01100 Yes None or Wait forthe kill to be successfully and SAD presented reset release reservation i pdate condition register store to cache mark cache block...

Страница 152: ...ondition register 010 011 I stwcx Write with 01M 10010 Yes None or Release reservation 010 flush and mm update condition register atomic reset store to main memory 011 I stwcx Write with 01M 10010 Yes...

Страница 153: ...10 n a None Load the block of data into cache mark the cache E 001 I debt Read 001 01010 nla m Load the block of data into cache mark the cache S 001 I debt Read 001 01010 nla 1J ITRYor Release the bu...

Страница 154: ...a Load the block of data into cache mark the cache S 000 I debtst Read 000 01010 Va or Release the bus retry the operation 000 s debtst None n a n a Va n a No op 000 ME debtst None 000 n a Va n a No o...

Страница 155: ...dcbz Kill 000 01100 n a None or Establish the block in data cache SRl5 without fetching the block from main memory clear all bytes mark cache block M 000 SI dcbz Kill 000 01100 n a AATRVor Release th...

Страница 156: ...rk cache block E 000 M dcbst Write with 100 00110 nla JJITFIVor Release the bus kill JJITFIV retry the operation 001 ESI debst Clean 001 00000 nla None or No op 001 ESI debst Clean 001 00000 nla ARTRY...

Страница 157: ...m retry the operation 000 I dcbf Flush 000 00100 n a None or No op SRI 000 I dcbf Flush 000 00100 n a ARTRY or Release the bus Am RV SR15 retry the operation 000 ES dcbf Flush 000 00100 nla None or Ma...

Страница 158: ...00 I debt Flush 100 00100 Va None or No op II 100 ES debt Flush 100 00100 Va None or Mark cache block I II 100 ESI debt Flush 100 00100 Va Am1fli or Release the bus Am1fli mti retry the operation 100...

Страница 159: ...the operation 001 EM debI Kill 001 01100 Va None or Mark cache block I m 001 EM debI Kill 001 01100 Va ARTRYor Release the bus AR TRV m retry the operation 011 I debI Kill W1M 01100 Va None or No op...

Страница 160: ...the operation 001 VAL lcbl ICBI 001 01101 nla None or Mark icache block INV fl 011 INV lcbl ICBI 01M 01101 nla None or No op 010 11M fl 110 111 011 INV lcbl ICBI 01M 01101 nla ARTRYor Release the bus...

Страница 161: ...rage invalidate mm instructions Wait for the completion of any outstanding storage instructions Invalidate the requested TLB entry Note This table does not thoroughly characterize the tlble instructio...

Страница 162: ...atomic E Snoop xx1 11010 wa Mark cache bbck S read atomic M Snoop xx1 11010 wa AJ tTRV Attempt to write cache block read backto main memory if atomic successful mark cache block S I Snoop xx1 01110 N...

Страница 163: ...write cache block RWITM and backto main memory atomic reset if successful markcache block I release reservation I Snoop xx1 00100 None None No op flush I Snoop xx1 00100 Yes None No op flush SE Snoop...

Страница 164: ...adox no one else should be write with and writing if this cache ls M flush reset Attempt to write cache block backto main memory if successful markcache block I release reservation I Snoop xx1 00110 N...

Страница 165: ...None Paradox no one else should be write with writing if this cache is E flush Mark cache block I atomic E Snoop xx1 10010 Yes None Paradox no one else should be write with and writing if this cache...

Страница 166: ...0 n a J FITRV Ha TLB invalidate Is pending SYNC respond with retry n a Snoop xx1 01001 n a None Hno TLB invalidates are TLBSYNC pending no op n a Snoop xx1 01001 n a J FITRV Ha TLB invalidate is pendi...

Страница 167: ...t and noncacheable direct store therefore the 604 does not maintain coherency for these operations and the cache is bypassed completely Memory forced direct store operations These operations are consi...

Страница 168: ...loating point status and control register FPSCR Additionally certain exception conditions can be explicitly enabled or disabled by software The PowerPC architecture requires that exceptions be taken i...

Страница 169: ...rmed by the software linked to the appropriate vector offset Exception handling is begun in supervisor level referredto as privilegedstate in the architecture specification Note that the PowerPC archi...

Страница 170: ...tion of a qualHied TEAindication on the 604 bus or the machine check input signal If the MSR ME is deared the processor enters the checkstop state when one of these signals is asserted Note that MSR M...

Страница 171: ...d exception condition is generated when either MSR FEO or MSR FE1 and FPSCR FEX are set The settings of FEO and FE1 are descrl ed in Table4 4 FPSCR FEX is set by the execution ol a floating point inst...

Страница 172: ...Reserved 01000 012FF Reserved for implementation specific exceptions not implemented on the 604 Instruction 01300 An instruction address breakpoint exception occurs when the address bits 0 to address...

Страница 173: ...where instruction execution should resume after the exception is handled When an exception occurs the address saved in machine status save restore register 0 SRRO is used to help calculate where inst...

Страница 174: ...rved bits are not saved Table 4 3 MSR Bit Settings Blt s Name Description 0 Reserved FuH Function 1 4 Reserved Partial function S 9 Reserved FuU function 10 12 Reserved Partial function 13 POW Power m...

Страница 175: ...branch instruction 23 FE1 IEEE floating point exception mode 1 See Table 4 4 24 Reserved This bit corresponds to the Al bit of the POWER architecture 25 IP Exception prefix The setting of this bit spe...

Страница 176: ...Bits FED FE1 Mode 0 0 Floating point exceptions disabled 0 1 Floating point imprecise nonrecoverable 1 0 Floatlng point Imprecise recoverable In the 604 this bit setting causes the 604 to operate in f...

Страница 177: ...reserved bits may not be copied 4 The MSR is set as describedin Table 4 3 The new values take effectbeginning with the fetching of the first instruction of the exception handler routine located at th...

Страница 178: ...er which they were issued The rfi instruction copies SRRl bits back into the MSR The instructions following this instruction execute in the context established by this instruction For a complete descr...

Страница 179: ...0 0 0 0 0 0 0 ILE Program 0 0 0 0 0 0 0 0 0 0 0 ILE Floating 0 0 0 0 0 0 0 0 0 0 0 ILE point unavaDable Decrementer 0 0 0 0 0 0 0 0 0 0 0 ILE System call 0 0 0 0 0 0 0 0 0 0 0 ILE Trace 0 0 0 0 0 0 0...

Страница 180: ...mot resume reliably the MSR RQ bit SRR1 30 is deared MSR POW 0 BE 0 ILE FE1 0 EE 0 IP PR 0 IA 0 FP 0 DR 0 ME RI 0 FEO 0 LE Set to value of ILE SE 0 The SRESET input provides a warm reset capability Th...

Страница 181: ...e bus operation If the MSR ME bit and the appropriate bits in HIDO are set the exception is recognized and handled otherwise the processor generates an internal checkstop condition When a processor is...

Страница 182: ...cution cannot resume in the same context that existed before the exception If the condition that caused the machine check does not otherwise prevent continued execution MSR ME is set to allow the proc...

Страница 183: ...and an attempt to fetch the next instruction fails This exception is implemented as it is defined by the PowerPC architecture OEA In addition an instruction fetch from a no execute segment results in...

Страница 184: ...olving double precision values aligned on a double word boundary require two accesses which are translated separately Ifeither translation creates a OSI exception condition that exception is signaled...

Страница 185: ...ode a floating point instruction that causes a floating point exception brings the machine to a precise state In doing so the 604 sequencer unit can detect floating point exception conditions and take...

Страница 186: ...e Programming Environments Manual When a system call exception is taken instruction execution resumes at offset OxOOCOO from the physical base address indicated by MSR IP 4 5 11 Trace Exception OxOODO...

Страница 187: ...in the MMCRO register As with other PowerPC interrupts the performance monitoring interrupt follows the normal PowerPC exception model with a defined exception vector offset OxOOFOO The priority of t...

Страница 188: ...serting the HALTED output regardless whether the clock is stopped Nap mode must be entered by using the following code sequence naploop sync mtmsr GPR modify the POW bit QD JJl at this point the EE bi...

Страница 189: ...the assertion of INT SRESET MCP or SMI when a decrementer interrupt occurs or when a hard reset is sensed For more information about the RUN and HALTED signals refer to Section 7 2 10 4 Run RUN lnput...

Страница 190: ...ed by PowerPC processors to locate the effective to physical address mapping for instruction and data accesses The segment information translates the effective address to an interim virtual address an...

Страница 191: ...56 Mbytes from the 32 bit effective address space into the physical memory space This canbe used for translating large address ranges whose mappings do not change frequently Direct store segments If t...

Страница 192: ...Mbyte sizes translation Implemented with IBAT and DBAT registers in BAT array Memory protection Architecturally defined Segments selectable as no execute Pages selectable as user supervisor and read o...

Страница 193: ...Bs with the following characteristics 128 entries two way set associative 64 x 2 LRU replacement Data TLB supports the DMMU instruction TLB supports the IMMU Hardware TLB update Hardware update of mem...

Страница 194: ...nd to enforce the protection hierarchy programmed by the operating system Section 4 3 Exception Processing describes the MSR which controls some of the critical functionality of the MMUs The figures s...

Страница 195: ...Data Instruction Accesses Accesses OI OI I 0 ifi ifi A20 A31 PAG PA31 Figure 5 1 MMU Conceptual Block Dlagram 32 Blt Implementations 5 6 PowerPC 604 RISC Microprocessor User s Manual...

Страница 196: ...Instruction Unit 1 PAO PA31 128 Sets l _a_ch_ _ 0 TAGS Select A20 A26 3 TAGS TAGS TAGS 0 19 Compare I Cache Hit Miss Figure 5 2 PowerPC 604 Microprocessor IMMU Block Diagram Chapter 5 Memory Manageme...

Страница 197: ...8 PAO PA31 128 Sets o _a_chre 0 TAGS Select A20 A26 3 0 TAGS TAGS TAGS 19 Compare DCache Hit Miss Figure 5 3 PowerPC 604 Microprocessor DMMU Block Diagram PowerPC 604 RISC Microprocessor Uaer s Manua...

Страница 198: ...the implementation invokes an elaborate hardware protocol for communication with these devices The direct store interface protocol is not optimized for performance and therefore its use is discourage...

Страница 199: ...aining information in the segment descriptor is interpreted as identifier information that is used with the remaining effective address bits to generate the packets used in a direct store interface ac...

Страница 200: ...no execute option is provided in the segment descriptor Each of the remaining options is enforced based on a combination of information in the segment descriptor and the page table entry Thus the supe...

Страница 201: ...ed bit or the changed bit is not set the hardware automatically sets both bits in the translation table In addition during the address translation portion of a store operation that hits in the TLB the...

Страница 202: ...n Real Addressing Mode and Block Note that if the BAT array search results in a hit the access is qualified with the appropriate protection bits If the access violates the protection mechanism an exce...

Страница 203: ...in Figure 5 6 In addition Figure 5 6 also shows the way in which the no execute protection is enforced if the N bit in the segment descriptor is set and the access is an instruction fetch the access...

Страница 204: ...are Virtual Address with TLB Entries L J TLB Miss TLB Hit _ See Figure 5 8 Perform Page Table Search Operation See Figure 5 9 PTENot Found PTE Found r _I_ Access Permitted Translate Address Continue A...

Страница 205: ...criptor has the T bit set the access is considered a direct store interface access and the direct store interface protocol of the external interface is used to perform the access to direct store space...

Страница 206: ...mory either matching xBAT G 1 or no matching BAT SRR1 3 1 entry and PTE G 1 In addition to the translation exceptions there are other MMU related conditions some of them defined as implementation spec...

Страница 207: ...ruction attempted while MSR LE 1 little endian mode Operand misalignment Translation enabled and operand is Alignment exception some misaligned as descrbed in Chapter 4 of these cases are Exceptions i...

Страница 208: ...PC 604 Processor Programming Model Table 5 6 PowerPC 604 Microprocessor MMU Registers Register Description Segment registers The sixteen 32 bit segment registers are present only in 32 bit Implementat...

Страница 209: ...hat is not subject to normal virtual memory handling paging such as a memory mapped display buffer or an extremely large array ofnumerical data Block address translation in the 604 is described in Cha...

Страница 210: ...f the page based on this access Table 5 7 Table Search Operations to Update History Blts TLB Hit case RandCbHs Processor Action lnTLB Entry 00 Combination doesn t occur 01 Combination doesn t occur 10...

Страница 211: ...of the PTE loaded into the TLB if a TLB is implemented as in the 604 Whenever a data store instruction is executed successfully if the TLB search for page address translation results in a hit the chan...

Страница 212: ...table load operations include those generated by load instructions by the eciwx instruction andby the cache managementinstructions that are treated as a load with respect to address translation Simila...

Страница 213: ...nt registers for the MMU the 604 maintains two identical sets of segment registers one for the IMMU and one for the DMMU when a segment register instruction executes the 604 automatically updates both...

Страница 214: ...e matching PTE is found in memory it is loaded into a particular TLB entry selected by the least recently used LRU replacement algorithm and the translation process begins again this time with a TLB h...

Страница 215: ...lidates four entries both the ITLB entries indexed by EA14 EA19 and both the indexed entries of the DTLB Execution of the tlbie instruction causes all entries in the congruence class corresponding to...

Страница 216: ...can be used to invalidate a particular index of the TLB based on EA 14 19 With that concept in mind a sequence of 64 tlbie instructions followed by a single tlbsync instruction would cause all the 604...

Страница 217: ...ception Check Page Memory Protection Violation Conditions see The Programming Environments Manualj Access Penniltad Access Prohibited see The Programming EnVlronmfltlts Manual PTE CJ 0 otherwise Page...

Страница 218: ...the following must be true PTE H 0 PTE V 1 PTE VSID VA 0 23 PTE API VA 24 29 4 If a match is not found step 3 is repeated for each of the other seven PTEs in the primary PTEG If a match is found the...

Страница 219: ...ip cache Figure 5 9 and Figure 5 10 show how the conceptual model for the primary and secondary page table search operations described in The Programming Environments Manual are realized in the 604 Fi...

Страница 220: ...E intoTLB otherwise Check Memory Protection VIOiation Conditions Aocess Permitted ocess Prohibited dcbz Instruction withWor1 1 R_ otherwlse PTE R 1 Update PTE R mMemory otherwl oreoperation with PTE C...

Страница 221: ...operation begins Once the matching PTE is found in memory it is loaded into the appropriate TLB entry depending on the LRU bit setting and translation continues The LSU initiates out of order accesse...

Страница 222: ...a PI E Deleting a PI E PI Es must be locked on multiprocessor systems Access to PI Es must be appropriately synchronized by software locking of that is guaranteeing exclusive access to PI Es or PI EG...

Страница 223: ...ring any higher level synchronization However extreme care must be taken to ensure that no store overwrites one of these bytes accidentally Processors write referenced and changed bits with unsynchron...

Страница 224: ...m these accesses involves the transfer ofaddress and data information in packets however the PowerPC OEA does not define the exact hardware protocol used for direct store interface accesses Some instr...

Страница 225: ...cbi 5 5 5 Direct Store Segment Translation Summary Flow Figure 5 11 shows the flow used by the MMU when direct store segment address translation is selected This figure expands the direct store segmen...

Страница 226: ...or stwcx instruction otherwise r L DSISR S 1 otherwise dcbtst dcbf dcbl dcbst L J dcbz or lcbl OSI Exception or Boundedly Undefined Results Perform Direct Store Interface Access _ C No Op Optional to...

Страница 227: ...e is forced to stall in its cycle In some cases an instruction may also occupy more than one stage simultaneously for example instructions may complete and write back their results in the same cycle A...

Страница 228: ...ions on the mispredictedpath are purged from the instructionpipeline and arereplaced with the instructions from the nonpredicted path Program order The original order in which program instructions are...

Страница 229: ...Overview The 604 has been designed to maximize instruction throughput and minimize average instruction execution latency For many ofthe instructions in the 604 this can be simplified to include only t...

Страница 230: ...ne instruction per clock cycle can be realized by the many performance features in the 604 including multiple execution units that operate independently and in parallel pipelining superscalar instruct...

Страница 231: ...uctions per cycle are completed in program order In the write back stage results are returned to the register file Instructions are fetched and executed concurrently with the execution and write back...

Страница 232: ...e pipelines The division of instructions into branch integer load store and ftoating poirit instructions indicates the execution unit in which the instructions execute For example mtspr instructions w...

Страница 233: ...ages Figure 6 4 PowerPC 604 Microprocessor Pipeline Stages Table 6 1 lists the latencies and throughputs for general groups of instructions Table 6 1 Execution Latencies and Throughputs Instruction La...

Страница 234: ...ation by the decode stage The instruction fetching logic is shown in Figure 6 5 BTAC Decode Buffer Decode Prediction Dispatch Buffer Dispatch Prediction Pending Branch Queue Target 1 Seq 1 BPU Reserva...

Страница 235: ...gic may indicate based on the BHT or an unconditional branch decode that an earlier BTAC prediction was incorrect The BPU can indicate that a previous branch prediction either from the BTAC or the dec...

Страница 236: ...e flushed and fetching resumes at the correct address If an instruction causes an exception the execution unit reports the exception to the complete stage and continues executing instructions regardl...

Страница 237: ...hose restrictions necessary to support a precise exception model the 604 imposes the following restrictions per each cycle Completion stops before a store since store data is read directly from GPRs o...

Страница 238: ...ck any information from the rename buffers that was not written back by the complete stage As mentioned in Section 6 2 1 1 5 Complete Stage each of the rename buffers has two read ports for write back...

Страница 239: ...4 Memory Operations briefly describes how these modes may affect instruction timing 6 3 1 MMU Overview The 604 implements separate 128 entry two way set associative TLBs one each for instruction and d...

Страница 240: ...without having to wait for the actual store operation to take place either in the cache or in system memory When the cache is not busy one completed store can be written to the cache per cycle In the...

Страница 241: ...as many as three outstanding pipelined operations The BIU can complete one or more write transactions between the address and data tenures of a read transaction The BIU provides critical double word f...

Страница 242: ...memory for example video memory or when there is shared global data that may be used frequently or when allocation of a cache block on a cache miss is undesirable Cached data is not automatically writ...

Страница 243: ...e operand becomes available The 604 contains the following execution units that operate independently and in parallel Branch processing unit BPU Two 32 bit single cycle integer units SCIU One 32 bit m...

Страница 244: ...the instruction being asked for is in the on chip cache cache hit or whether a memory transaction is required to bring the data into the cache cache miss 6 4 2 1 Cache Hit Timing Example Assuming tha...

Страница 245: ...cycle 1 the last two instructions in the cache block addc and subfc are fetched while instructions 0 3 pass into the decode stage 2 In cycle 2 the two integer add instructions 0 and 1 are dispatched o...

Страница 246: ...structions were ready for dispatch Note that because ofin order dispatch the integer instructions 8 and 9 are also held in the dispatch stage behind the fmsub instruction The final pair of floating po...

Страница 247: ...ions move into execute stage along with the previous fsubs instruction which is in the first stage of execute The fmadd instruction completes and writes back and the subsequent floating point instruct...

Страница 248: ...hat this example asswnes a best case scenario 0 I I Io add I 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 I I I I I I I I Address I t 5add I I 1 l61subl h I I l7addl lllll l C J Fetch Execute llmlllllllllll 1...

Страница 249: ...bus and access has been granted for the data bus 6 In cycle 6 fadd 1 completes and writes back allowing the add 2 instruction to complete and write back The fadd 3 instruction is in the final execute...

Страница 250: ...be taken The 64 entry BTAC is fully associative to provide a high hit percentage Ifa fetch address is in the BTAC the target address is used in the next cycle to fetch the instructions from the predic...

Страница 251: ...xamples use the following code sequence and ld add be or cmp ld mulli 6 4 4 1 1 Timing Example Branch Timing for a BTAC Hit Figure 6 9 shows the timing for a branch instruction that had a BTAC hit 0 0...

Страница 252: ...br instruction also completes Because the branch is taken the or 4 instruction which could otherwise write back in this cycle stays in the complete stage and completes and writes backinthe nextcycle T...

Страница 253: ...execute stage of the LSU pipeline Instructions 2 and 3 wait in the complete stage Instructions 4 7 enter the dispatch stage 3 In cycle 5 the Id 1 instruction is able to write back allowing the follow...

Страница 254: ...1 Execute IIIIIII1 Complete Write Back 4 5 6 7 8 I I I I I I I I I 1111 Figure 6 11 Instruction Timing Branch with BTAC Miss Dispatch Correction 6 4 4 1 4 Timing Example Branch with BTAC Miss Execute...

Страница 255: ...To record the predicted state the 604 uses many of the same resources primarily the rename buffers and completion buffer and logic as the mechanism used to maintain a precise exception model as is co...

Страница 256: ...h have two entry in order reservation stations These stations allow instructions to clear the dispatch stage even though operands may not yetbe available for execution to occur The BPU FPU and LSU ins...

Страница 257: ...name registers are provided for the GPRs eight for the FPRs and eight for the condition register A GPR rename buffer entry is allocated when an instruction that modifies a GPR is dispatched This entry...

Страница 258: ...GPR has not yet executed In this case the instruction is dispatched with the rename buffer entry identifier inplace ofthe operand which will be supplied by the reservation station when the result is p...

Страница 259: ...to that unit to fill the remaining pipeline stages 6 4 7 Instruction Serialization Some instructions such as mfspr and most mtspr instructions extended arithmetic instructions that require the carry...

Страница 260: ...until the completion block informs the execution unit to execute the instruction This means it is guaranteed to wait at least one cycle before it can execute Instructions causing execution serializat...

Страница 261: ...nch Unit Instruction Timings The 604 can have two unresolved branches in the branch reservation station and two resolved branches that have not yet completed The branch unit serves to validate branch...

Страница 262: ...ervation Station I Rotate Shift i Adder Logic I CTLZ Comparator ics j_ 3 1 MUX I j Figure 14 SCIU Block Diagram The MCIU which handles all integer multiple cycle integer insttuctions consists of a 32...

Страница 263: ...are not execution serialized 6 5 3 Floating Point Unit Instruction Timings The floating point unit on the 604 executes all floating point instructions Execution of most floating point instructions is...

Страница 264: ...zed data types Instructions are obtained from the instruction dispatcher and placed in the reservation station queue The operand sources are the FPR the floating point rename buffers and the result bu...

Страница 265: ...of order execution are resolved by the LSU The se dependencies arise when in the instruction stream a store is followed by a load from the same address If the load instruction is speculatively execute...

Страница 266: ...cuted until all of the instruction operands are valid The address calculation block includes a 32 bit adder that computes the effective address for all operations The data alignment blocks manage the...

Страница 267: ...branch unit instructions as specified by the PowerPC architecture but they do not actually execute in the BPU in the same sense that other branch instructions do The completion unit treats the rfi and...

Страница 268: ...MCIU executes all integer multiply divide and move to from instructions exceptmtcrfinstructions thatupdate only one field which areexecutedin either of the SCIUs The load store unit executes load stor...

Страница 269: ...in the execute andcomplete stages The dispatchstalls onthe first instruction after the fourth branchuntil the first branch completes An instruction cannot be dispatched until all destination register...

Страница 270: ...hen the option exists Forexample an integer register copy can be performed in a single cycle using a number of different instructions However using an ori instruction with an immediate operandofzero u...

Страница 271: ...ching nevertheless changing control flow in aprogram is relatively expensive in that fullest advantage cannot be taken ofresources that can improve throughput such as superscalar instruction dispatch...

Страница 272: ...eline Integer multiplies other than those that can exit early described in the previous bullet stall for one cycle in the first stage ofthe pipeline Integerdivide instructions iterate in stage two oft...

Страница 273: ...cntlzw SCIU 1 crand BPU 1 Execute crandc BPU 1 Execute creqv BPU 1 Execute crnand BPU 1 Execute crnor BPU 1 Execute cror BPU 1 Execute crorc BPU 1 Execute crxor BPU 1 Execute dcbf LSU Execute dcbl LSU...

Страница 274: ...ctlwz FPU 3 fdlv FPU 32 FPeq ty1 tdlvs FPU 18 FPeq ty1 tmadd FPU 3 tmadds FPU 3 tmr FPU 3 tmsub FPU 3 tmsubs FPU 3 tmul FPU 3 tmuls FPU 3 tnabs FPU 3 tneg FPU 3 tnmadd FPU 3 tnmadds FPU 3 tnmaub FPU 3...

Страница 275: ...LSU 3 lfsu LSU 3 lfsux LSU 3 lfsx LSU 3 Iha LSU 2 lhau LSU 2 lhaux LSU 2 lhax LSU 2 lhbrx LSU 2 lhz LSU 2 lhzu LSU 2 lhzux LSU 2 lhzx LSU 2 lmw LSU regs 2 String multiple lswl LSU 2 regs 2 String mult...

Страница 276: ...MCIU 1 DispatclVExecute mtcrf single bit SCIU 1 mtfsbO FPU 3 mtfsb1 FPU 3 mtfsf FPU 3 mtfsfl FPU 3 mtmsr MCIU 1 Execute mtspr LR CTR MCIU 1 Dispatch mtspr XER MCIU 1 Complete 2 mtspr others MCIU 1 Ex...

Страница 277: ...dux LSU 3 Execute stfdx LSU 3 Execute stflwx LSU 3 Execute SUS LSU 3 Execute stfsu LSU 3 Execute stfsux LSU 3 Execute stfsx LSU 3 Execute sth LSU 3 Execute sthbrx LSU 3 Execute sthu LSU 3 Execute sthu...

Страница 278: ...U tlbl LSU Execute tlbaync LSU tw SCIU 1 twl SCIU 1 xor SCIU 1 xorl SCIU 1 xorl SCIU 1 1 These Instructions ara not pJpeOned They cannol be executed untl the pnwlous lnstrucllon In the FPU completes s...

Страница 279: ...ss transfer start signals These signals indicate that a bus master has begun a transaction on the address bus Address transfer signals These signals which consist of the address bus address parity and...

Страница 280: ...rupt and under various conditions to reset the processor JTAG COP interface signals The JTAG IEEE 1149 1 interface and common on chip processor COP unit provides a serial interface to the system for p...

Страница 281: ...OP INTERFACE Figure 7 1 PowerPC 604 NJlcroprocessor Signal Groups 7 2 Signal Descriptions This section describes individual 604 signals grouped according to Figure 7 1 Note that the following sections...

Страница 282: ...aster that assertedARTRY due to the need to perform a cache line push 7 2 1 2 Bus Grant lm lnput The bus grant BG signal is an input signal on the 604 Following are the state meaning and timing commen...

Страница 283: ...cycle following a qualified bus grant the 604 did not accept mastership even if BR was asserted High Impedance Occurs one half bus cycle two thirds bus cycle when using 3 1 clock mode and one third b...

Страница 284: ...s Assertion Coincides with the assertion ofABB Negation Occurs one bus clock cycle after TS is asserted High Impedance Occurs one bus clock cycle after TS is negated 7 2 2 1 2 Transfer Start TS lnput...

Страница 285: ...ls 7 2 3 1 1 Address Bus AO A31 0utput Memory Operations Following are the state meaning and timing comments for the AO A31 output signals State Meaning Asserted Negated Represents the physical addres...

Страница 286: ...timing comments for input direct store operations on the 604 State Meaning Asserted Negated When the 604 is not the master it snoops and checks address parity onthe first address beatonly ofall direct...

Страница 287: ...r deassertion of the APE signal Following are the state meaning and timing comments for the APE signal on the 604 For more information see Section 8 3 2 1 Address Bus Parity State Meaning Asserted Ind...

Страница 288: ...sfer in progress see Table 7 1 For direct store operations the TTO TT3 signals form part of the XATC and are snooped by the 604 ifXATS is asserted Timing Comments Assertion Negation The same as AO A31...

Страница 289: ...beat Caching inhibited read or burst load 0 1 1 1 0 Read with Intent Burst Load miss or store to modify miss 1 0 0 1 0 Write with flush Single beat stwcx atomic write 1 0 1 1 0 Reserved NIA NIA 1 1 0...

Страница 290: ...nd ecowx TSIZO TSIZ2 are used to output bits 29 31 of the external access register EAR which are used to form the resource ID TBSTllTSIZO TSIZ2 Timing Comments Assertion Negation The same as AO A31 Hi...

Страница 291: ...eaning Asserted Negated For the 1 0 transfer protocol this signal forms part ofthe 1 0 transfer code see Section 7 2 4 1 TransferType TTO TT4 Timing Comments Assertion Negation The same as AO A31 7 2...

Страница 292: ...ordinary and atomic read and read with intent to modify operations 2 ICBI operation is distinguished from kiU block by assertion of TT4 bit 3 W write through bit from translation The value shown in t...

Страница 293: ...is not global Timing Comments Assertion Negation The same as AO A31 High Impedance The same as AO A31 7 2 4 7 2 Global GBL lnput Following are the state meaning and timing comments for the GBL input s...

Страница 294: ...for slow devices For example if an implementation supports slow snooping devices an external arbiter can postpone the assertion ofAACK Negation Must occur one bus clock cycle after the assertion of AA...

Страница 295: ...bus master to allow an opportunity for a copy back operation to main memory Note that the subsequent address presented on the address bus may not be the same one associated with the assertion of the A...

Страница 296: ...tailed information about using DBWO see Section 8 11 Using Data Bus Write Only 7 2 6 1 Data Bus Grant DBG lnput The data bus grant DBG signal is an input signal input only on the 604 Following are the...

Страница 297: ...DBB output signal State Meaning Asserted Indicates that the 604 is the data bus master The 604 always assumes data bus mastership if it needs the data bus and is given a qualified data bus grant see D...

Страница 298: ...is there are no 64 bit 1 0 transfers Timing Comments The data bus is driven once for noncached transactions and four times for cache transactions bursts Table 7 4 Data Bus Lane Assignments Data Bus Si...

Страница 299: ...ts are listed in Table 7 5 Timing Comments Assertion Negation The same as DLO DL31 High Impedance The same as DLO DL31 Table 7 5 DPO DP7 Slgnal Assignments Signal Name Signal A slgnments DPO DHO DH7 D...

Страница 300: ...tate Meaning Asserted Indicates for a write transaction that the 604 must release data bus and the data bus parity to high impedance during the following cycle The data tenure will remain active DBB w...

Страница 301: ...or fast L2 mode must not be asserted earlier than the first cycle of avalid ARTRY window otherwise assertion may occur at any time during the assertion ofDBB The system can withhold assertion of TAto...

Страница 302: ...e MSR ME 0 Assertion terminates the current transaction that is assertion ofTA and DRTRY are ignored The assertion ofTEAcauses the negation high impedance ofDBB in the next clock cycle However data en...

Страница 303: ...d and negated synchronously with the SYSCLK signal 7 2 9 2 System Management Interrupt SMl lnput The system management interrupt SMI signal is input only Following are the state meaning and timing com...

Страница 304: ...llowing are the state meaning and timing comments for the CKSTP_IN signal State Meaning Asserted Indicates that the 604 must terminate operation by internally gating off all clocks and release all out...

Страница 305: ...ing Comments Assertion May occur at any time and may be asserted asynchronously to the 604 input clock must be held asserted for a minimum of 255 clock cycles Negation May occur any time after the min...

Страница 306: ...ng Comments Assertion Negation May occur on any cycle 7 2 10 2 Reservation RSRV Output The reservation RSRV signal is output only on the 604 Following are the state meaning and timing comments for the...

Страница 307: ...e to the 604 entering nap mode or a JTAG COP request Negated Indicates that internal clocks are running Timing Comments Assertion Negation Occurs synchronously with internal processor clock For additi...

Страница 308: ...bus interface Internally the 604 uses a phase lock loop PLL circuit to generate a master clock for all of the CPU circuitry including the bus interface circuitry which is phase lockedto the SYSCLK inp...

Страница 309: ...ion 7 2 12 4 PLL Configuration PLL_CFGO PLL_CFG3 lnput The PLL phase lock loop is configured by the PLL_CFGO PLL_CFG3_pins For a given SYSCLK bus frequency the PLL configuration pins set the internal...

Страница 310: ...04 hardware specifications for tlmilg comments PLL frequencies shown In parenthesis In the table above should not fall below 100 MHz and should not exceed 200 MHz 2 In PLL bypass mode the SVSCLK i11 u...

Страница 311: ...tions are automatically fetched from the memory system into the instruction unit where they are dispatched to the execution units at apeakrate offour instructions per clock Conversely load and store i...

Страница 312: ...ough memory read and write operations Additionally there can be address only operations variants of the burst and single beat operations global memory operations that are sno9ped and atomic memory ope...

Страница 313: ...NIT 16 Entry Reorder Buffer 32Bit Fetcher Instruction Queue 8 word GPRFile Rename Buffers 12 rn 128Bit DMMU 64Bit Store Queue Finish Load I Ilo ATI IDTLB I array i 1_ _ 1 Queue 32Bit 32 BIT ADDRESS BU...

Страница 314: ...is flexible allowing the 604 to be integrated into systems that implement various fairness and bus parking procedures to avoid arbitration overhead Typically memory accesses are weakly ordered sequen...

Страница 315: ...ing Diagram Legend 8 1 3 Direct Store Accesses Memory and direct store accesses use the 604 signals differently The 604 defines separate memory and 1 0 address spaces or segments distinguished by the...

Страница 316: ...that the address and data tenures are distinct from one another and that both consist of three phases arbitration transfer and tennination Address and data tenures are independent indicated in Figure...

Страница 317: ...ly the 604 s retry capability provides an efficient snooping protocol for systems with multiple memory systems including caches that must remain coherenL 8 2 1 Arbitration Signals Arbitration for both...

Страница 318: ...Dim data bus busy Assertion by the 604 indicates that the 604 is the data bus master The 604 always assumes data bus mastership if it needs the data bus and is given a qualified data bus grant see DBG...

Страница 319: ...astership of the address bus to the next requesting master before the current data bus tenure has completed Three address tenures can occur before the current data bus tenure completes The 604 can pip...

Страница 320: ...e previous cycle the 604 has what is referred to as a qualified bus grant The 604 assumes address bus mastership by asserting ABB when it receives a qualified bus grant Logical Bus Clock ne8d_bus 1 Im...

Страница 321: ...sserting ABB and negating the BR output signal Meanwhile the 604 drives the address for the requested access onto the address bus and asserts TS to indicate the start of a new transaction When designi...

Страница 322: ...e T TT4 transfer code TCO TC2 transfer size TSIZO TSIZ2 transfer burst TBST cache inhibit Cl write through WT global GBL and cache set element CSEO CSEl Figure 8 6 shows that the timing for all of the...

Страница 323: ...r Attribute Signals The transfer attribute signals include several encoded signals such as the transfer type TTO TT4 signals transfer burst TBST signal transfer size TSIZO TSIZ2 signals and transfer c...

Страница 324: ...e cache in order Burst write transfers are always performed zero double word first butsince burst reads are performedcritical double word first a burst read transfermay not start with the first double...

Страница 325: ...nerate burst memory operations if they miss the 604 interface supports misaligned transfers within a word 32 bit aligned boundary as shown in Table 8 5 Note that the four byte transfer in Table 8 5 is...

Страница 326: ...th a size of fewer than four bytes For the first bus operation bits A29 A31 equals bits 29 31 of the data which will be OblOl ObllO or Oblll The size associated with the first bus operation will be 3...

Страница 327: ...can be used with the WT TTO TI4 and TBST signals to further define the current transaction When asserted the transfer codes have the following meanings TCO Read cycle indicates code fetch Write cycle...

Страница 328: ...ress transfer until the AACK address acknowledge input is asserted therefore the system can extend the address transfer phase by delaying the assertion of AACK to the 604 AACK can be asserted as early...

Страница 329: ...le following the first or only assertion of TA for the data tenure the 604 ignores the first data beat and if it is a load operation does not forward data internally to the cache and execution units I...

Страница 330: ...re 8 7 H the data bus is needed the arbiter grants data bus mastership by asserting the DBG input to the 604 As with the address bus arbitration phase the 604 must qualify the DBG input with a number...

Страница 331: ...e cycles before the beginning of the snoop window in fast L2 data streaming then data is transferred too early to be cancelled by ARTRY Therefore systems in which ARTRY can be asserted must not attemp...

Страница 332: ...a previously requested read data tenure Following the ARTRY assertion the qualified DBG assertion to the 604 in cycle 7 will be accepted for the read data tenure 2 3 4 s I 6 I 7 8 9 10 System Clock TS...

Страница 333: ...ddress tenures were performed The 604 however also supports a limited out of order capability with the data bus write only DBWO input The DBWO capability exists to alleviate deadlock conditions that a...

Страница 334: ...all other burst operations however the Cil he line write operations are transferred beginning with the oct word aligned data and burst reads begin on double word boundaries The 604 does not directly s...

Страница 335: ...ses the data tenure to be tenninated immediately ifthe AR IRY is for the address tenure associated with the data tenure in operation the data tenure may not be tenninated due to address pipelining If...

Страница 336: ...ure 8 12 The bus clock cycles in which TA is asserted need not be consecutive thus allowing pacing of the data transfer beats For read bursts to terminate successfully TEAand DRlRY must remain negated...

Страница 337: ...in Figure 8 13 DBB cannot be asserted until bus clock cycle 5 This is true for both read and write operations even though DRTRY does not extend bus mastership for write operations 2 3 mm I M Jt data I...

Страница 338: ...g the TEA so that the bus operations associated with a machine check exception can proceed Due to bus pipelining in the 604 all outstanding bus operations including all queued requests are completed i...

Страница 339: ...t result in the generation of machine check exceptions Note that TEA generates a machine check exception depending on the ME bit in the MSR Clearing the machine check exception enable control bit lead...

Страница 340: ...01 Note that write hits to clean lines ofnonglobal pages do not generate invalidate broadcasts There are several types of bus transactions that involve the movement of data that can no longer access t...

Страница 341: ...Intent to Modify SHR Snoop Hit on a Read SHW Snoop Hit on a Write or D Cache Block Fill Read with Intent to Modify Figure 8 15 MESI Cache Coherency Protocol State Diagram WIM 001 Table 8 7 shows the...

Страница 342: ...e stated between bus tenures 8 32 1112 31415 6 I 7 I a I 9 I 10 11 12 BR l _ W I _ _ __ __ I 1 m A 1 b A 1 1 TPn I t 1 1 l DD I i ____ i ___ J T I I I I I AO A311 I l _ c_ru_A c Pu_A 1 acru jj AA l l...

Страница 343: ...1 I 1 I 2 a I 4 I s I 6 7 I s I 9 I 10 11 I 12 BR _ _ _ _ _ _ _ _ ____ 1__ _ _ _ _ ___ mi lffi lffi Mlf W ABB I I r I 1 ___ I 1 r _ I I TS I I I AO A31 1 f CPiiA I I _ _ I TIO TI4 I sew sew sew I I I...

Страница 344: ...three stated between bus tenures 1112 31415 6 I 7 I s 9 I 10 I 11 I 12 I 13 I 14 I BR1 I I UG i t w s t m i l t w f i __ _1 J 4 At F mu ABBI I I I_ I I I I 1 1 j I 1 1 i TS I I I I A0 A31 I 4 CPU A J...

Страница 345: ...2 m __ __ r I _ T r 1 zj h 1 4 xim I TS I _ ___ __ rt ___ I I I I AO A31 1 1 r cFiii A l __ CPU A J l I I I I I I TTO TT4 I I saw saw H r siBiawN t 4 ll I I I I I I I I I I i i i i i i i i i i i i i H...

Страница 346: ...sfer is delayed until the first transfer completes I 1 I 2 I a I4 Is I s I 1 I s I 9 110111 112 11a 11411s 11s 11111s 119 120 I AO A31 I TTO TT4 8 36 I TBSTI I I I q I XRTRV l 1 I t5DCI f5 b j j j i A...

Страница 347: ...I 9 11011q1211a 11411s 11s 11r 1 tmt I l J I I I I I I I Em r I tJ I I BF r T r r I I I I I I ABS I I I __ _ I I I TS1 I I I I I AO A31 bF lTA 1 r CPITA _ ___ TIO TI4 TBSTI I I I I I I I Gm t t I h H...

Страница 348: ...segment accesses The direct store protocol for the 604 allows for the transfer of 1 to 128 bytes of data between the 604 and the bus unit controller BUC for each single load or store request issued b...

Страница 349: ...ng 604 transactions effectively indicating to the 604 that the BUC is in a queue full condition and cannot accept new data In addition to the extensions noted above there are fundamental differences b...

Страница 350: ...wo or more direct store operations for loads from the 604 and one reply operation from the addressed BUC Table 8 8 Direct Store Bus Operations Operation Address Only Direction XATC Encoding Load start...

Страница 351: ...re reply operation no store immediate operation is involved in the transfer as shown in the following sequence STORE LAST from 604 STORE REPLY from BUC However ifmore data is involved in the direct st...

Страница 352: ...ee of the seven defined operations are address only transactions and do not use the data bus However unlike the memory transfer protocol these transactions are not broadcast from one master to all sno...

Страница 353: ...ent operating context user or supervisor Note thatuser and supervisor level refer to problem and privileged state respectively in the architecture specification Segment register Address bits 3 27 corr...

Страница 354: ...egister bits 28 31 II effective address bits 4 31 While the 604 provides the address of the transaction to the BUC the BUC must maintain a valid address pointer for the reply 8 6 3 1 0 Reply Operation...

Страница 355: ...he bus and on reply operations compares this field to bits 28 31 of the PIO register to determine HIt should recognize this VO reply The second beat of the address bus is reserved the XATC and address...

Страница 356: ...y as 128 bytes ofdata in one load or store instruction requiring more than 33 immediate operations in the case ofmisaligned operands In Figure 8 26 XATS is asserted with the same timing relationship a...

Страница 357: ...direct store operation indicates that an unrecoverable error has occurred If the TEA signal is asserted during a direct store operation the TEA action will be delayed and following direct store trans...

Страница 358: ...han in the normal bus protocol The PowerPC bus protocol specifies that during load operations the memory system normally has the capability to cancel data that was read by the master on the bus cycle...

Страница 359: ...he basic and extended transfer bus protocols described in this chapter 8 7 1 1 Fast L2 Data Streaming Mode Design Considerations It is recommended that use of fast L2 data streaming mode be accompanie...

Страница 360: ...ted AR1RY signal invalidates the data that was transferred the previous cycle in the same way DRTRY cancels data from the previous cycle In fast L2 data streaming mode the data buffering that allows l...

Страница 361: ...oft reset The soft reset input provides warmreset capability This input can be used to avoid forcing the 604 to complete the cold start sequence When either reset input is negated the processor attemp...

Страница 362: ...ant implementation of the IEEE 1149 1 standard This section describes the 604 IEEE 1149 l JTAG interface 8 10 1 IEEE 1149 1 Interface Description The 604 has five dedicated JTAG signals which are desc...

Страница 363: ...een performed because of address pipelining The DBWO does not change the order of write tenures with respect to other write tenures from the same 604 It only allows that a write tenure be performed ah...

Страница 364: ...DBWO Any number of bus transactions by other bus masters can be attempted between any of these steps Note the following regarding DBWO The DBWO signal can be asserted ifno data bus read is pending but...

Страница 365: ...ot easily be characterized by a benchmark or trace To help system developers bring up and debug their systems The perfonnance monitoruses the following 604 specific special purpose registers SPRs Perf...

Страница 366: ...ed instruction address SIA register and the sampled data address SDA register respectively For more infonnation see Section 9 1 2 2 Threshold Events For all other programmable events that cause a PMI...

Страница 367: ...sing the mfspr and mtspr instructions Software is expected to use the mtspr instruction to explicitly set the PMC register to nonnegative values If software sets a negative value an erroneous interrup...

Страница 368: ...store data cache misses that exceeded the threshold value with lateral L2 cache intervention 0001011 Number of mtspr instructions dispatched 0001100 Number of sync instructions completed 000 1101 Num...

Страница 369: ...completed every cycle 00 None 01 Illegal value 10 One 11 Two 001001 Number of reservations successfully obtained 001010 Number of mfspr instructions dispatched speculative 00 1011 Number of lcbl inst...

Страница 370: ...processor signals the performance monitor interrupt condition In this case the SDA is not meant to have any connection with the value in the SIA If the performance monitor interrupt was triggered by a...

Страница 371: ...can be changed by hardware 1 HMSR PM is set the PMCn counters are not changed by hardware 4 OMA Disable counting while MSR PMJ is zero 0 The PMCn counters can be changed by hardware 1 nMSR PM is clear...

Страница 372: ...r interrupt is signaled 0 Enable PMC2 counting 1 Disable PMC2 counting until PMC1 bit OIs set or unto aperformance monitor interrupt is signaled This signal can be used to trigger counting of PMC2 aft...

Страница 373: ...instruction The 32 bit registers can count up to Ox7FFFFFFF 2 147 483 648 in decimal before becoming negative The most significant bit bit 0 of both registers is used to determine if an interrupt cond...

Страница 374: ...struction of a load miss distribution chart When a load or store miss arrives in the load store queue the threshold control logic begins decrementing For each cycle that passes the threshold value in...

Страница 375: ...provide this functionality the events that use this signal PMCl events 9 and 10 become obsolete 9 1 2 3 Nonthreshold Events Nonthreshold events are all events except for PMCl events 9 10 23 or 24 Any...

Страница 376: ...n lowercase For more information refer to Chapter 8 Instruction Set in The Programming Environments Manual A 1 Instructions Sorted by Mnemonic Table A 1 lists the instructions implemented in the 604 i...

Страница 377: ...9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 s A UIMM s A UIMM LI BO Bl BO BO Bl 528 LK BO Bl 16 LK crfD if L A B 0 crfD L A SIMM crfD if L A B 32 crfD ML A UIMM 26 crbD crbA c...

Страница 378: ...63 D A B c 29 Re fmaddsx 59 D A B c 29 Re fmrx 63 D B 72 Re fmsubx 63 D A B c 28 Re fmsubsx 59 D A B c 28 Re fmulx 63 D A c 25 Re fmulsx 59 D A c 25 Re fnabsx 63 D B 136 Re fnegx 63 Re D 40 B fnmaddx...

Страница 379: ...8 49 31 31 42 43 31 31 31 40 41 31 31 46 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 d D A d D A d D A d D A B 631 tf j D A B 599 i i i D A d D A d D A B 567 I D A B 535...

Страница 380: ...f mtfsbOx mtfsb1x mtfsfx mtfsflx mtmsr 1 mtspr 2 mtsr 1 31 D A NB 31 D A B 32 D A 33 D A 31 D A B 31 D A B 19 63 31 31 63 31 31 D spr 31 D SR 31 Q j d tl ii B D 31 D tbr 31 63 63 63 63 31 31 S spr 31...

Страница 381: ...mullwx 31 D A B 235 Re nandx 31 s A B 476 Re negx 31 D A rn RR 104 Re norx 31 s A B 124 Re orx 31 s A B 444 Re orcx 31 s A B 412 Re Ori 24 s A UIMM orls 25 S A UIMM rlwlmlx 20 s A SH MB ME Re rlwlnmx...

Страница 382: ...d slfsu 53 s A d stfsux 31 s A B 695 slfsx 31 s A B 663 sth 44 s A d sthbrx 31 s A B 918 t sthu 45 s A d sthux 31 s A B 439 Jt sthx 31 s A B 407 w stmw 3 47 s A d stswl 3 31 s A NB 725 stswx 3 31 s A...

Страница 383: ...08 D A 8 twl 03 TO xorx 31 s xort 26 s xorls 27 s 1 Supervisor level instruction 2Supervisor and user level instrudion 3Load and store siring or multiple instrudion 464 bit instrudion 5Optional instru...

Страница 384: ...A SIMM subtle 001000 D A SIMM cmpll 001010 A UIMM cmpl 001011 A SIMM addle 001100 D A SIMM addle 001101 D A SIMM addl 001110 D A SIMM addls 001111 D A SIMM bcx 010000 SC 010001 bx 010010 mcrf 010011 0...

Страница 385: ...100 andls 011101 cmpl 011111 subfx 011111 A 10 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 BO s s s s s s s s s crlD i L D Bl A A A A A A A A A A A SH SH B B B 1000010000 M...

Страница 386: ...0 1 0 1 1 1 negx 0 1 1 1 1 1 D A r i 000 1 1 0 1 000 Re lbzux 0 1 1 1 1 1 D A B 000 1 1 1 0 1 1 1 norx 0 1 1 1 1 1 s A B 000 1 1 1 1 1 00 Re subfex 0 1 1 1 1 1 D A B 00 1 000 1 000 Re addex 0 1 1 1 1...

Страница 387: ...1 iii addx 011111 D A B 0100001010 Re debt 011111 A B 01000101 1 0 lii lhzx 011111 D A B 01000101 1 1 iM eqvx 011111 s A B 010001 1 100 Re eclwx 011111 B 01001101 1 0 lhzux 011111 D A B 01001 101 1 1...

Страница 388: ...10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 D D D s D D D D D D s s s s s s s D s s s s s A A A A A SR A A A t t A _ _ rr r A A A A A A A A A A A A A A B 1000010101 B 1000010110...

Страница 389: ...d stbu 100111 s A d lhz 101000 D A d lhzu 101001 D A d Iha 101010 D A d lhau 101011 D A d sth 101100 s A d sthu 101101 s A d lmw 3 101110 D A d stmw 3 101111 s A d Ifs 110000 D A d lfsu 110001 D A d...

Страница 390: ...0000001100 Re fctlwx 111111 D B 0000001110 fctlwzx 111111 D B 0000001111 Re fdlvx 111111 fsubx 111111 faddx 111111 fselx 5 111111 D A B C 1 0 1 1 1 Re fmulx 111111 lrsqrtex 5 111111 fmsubx 111111 D A...

Страница 391: ...sx 111111 D fabsx mffsx 111111 111111 D A 16 1 Supervisor level instruction 2 Supervisor and user level instruction 3 Load and store string or multiple instruction 4 64 blt Instruction 5 Optional inst...

Страница 392: ...e 604 Table A 3 Integer Arithmetic Instructions Name 0 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 addx 31 D A B 266 Re addex 31 D A B 10 Re addex 31 D A B 138 Re addI...

Страница 393: ...L A UIMM Table A 5 Integer Logical Instructions Name o 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 andx 31 s A B 28 Re andcx 31 s A B 60 Re andI 28 s A UIMM andls 29 s...

Страница 394: ...ame o 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Table A 8 Floating Point Arithmetic Instructions Name 0 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 2...

Страница 395: ...30 31 D A B c 29 Re D A B c 29 Re D A B c 28 Re D A B c 28 Re D A B c 31 Re D A B c 31 Re D A B c 30 Re D A B c 30 Re Table A 10 Floating Point Rounding and Conversion Instructions Name o 5 6 7 8 9 10...

Страница 396: ...38 R mtfsfx 31 711 Re mtfsflx 63 134 Re Table A 13 Integer Load Instructions Name o 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 lbz lbzu lbzux 34 35 31 D A D A D A lha...

Страница 397: ...stwux 31 s A stwx 31 s A Table A 15 Integer Load and Store with Byte Reverse Instructions Name 0 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 lhbrx 31 D A B 790 1 lwbrx...

Страница 398: ...28 29 30 31 Table A 19 Floatlng Polnt Load Instructions Name 0 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 lfd 50 D A d lfdu 51 D A d lldux 31 D A B 631 I lfdx 31 D A...

Страница 399: ...tructions Name o 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 bx bcx bcctrx bclrx Name crand crandc creqv crnand cmor cror crorc crxor mcrf A 24 0 18 LI 16 BO Bl BO 19 B...

Страница 400: ...ions Name 0 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 31 512 mcrxr mfcr 31 D 19 mfmsr 1 31 D 83 mfspr 2 31 D 339 mftb 31 D 371 mtcrf 31 s 144 1 k mtmsr 1 31 s 146 I m...

Страница 401: ...uctions Name o 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Table A 30 External Control Instructions Name o 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26...

Страница 402: ...30 31 bxl 18 LI MLKI Table A 32 B Form OPCO BO Bl BO l viLKI Specific Instruction Name 0 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 bcxl 16 BO Bl BO MLKI Table A 33 S...

Страница 403: ...46 32 33 7 24 25 38 39 54 55 52 53 44 45 47 Specific Instructions 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 D A SIMM D A SIMM D A SIMM D A SIMM s A UIMM s A UIMM crtD...

Страница 404: ...ls 27 s A UIMM Table A 35 OS Form OPCD D A ds I XO I OPCD s A ds XO Specific Instructions Name 0 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Table A 36 X Form OPCD D A...

Страница 405: ...4 25 26 27 28 29 30 31 andx 31 s A B 28 Re andcx 31 s A B 60 Re cmp 31 crfD A B 0 lll dcbf 31 A B 86 1 debI 1 31 A B 470 il debst 31 A B 54 ill debt 31 A B 278 t l dcbtst 31 A B 246 llill dcbz 31 A B...

Страница 406: ...t B 136 Re fnegx 63 D Mm B 40 Re frspx 63 D Ji i MF t B 12 Re lebl 31 r g MMI ir A B 982 if lbzux 31 D A B 119 lbzx 31 D A B 87 lfdux 31 D A B 631 lfdx 31 D A B 599 1 lfsux 31 D A B 567 lfsx 31 D A B...

Страница 407: ...3 mcrfs 63 64 mcrxr 31 512 mfcr 31 19 mffsx 63 583 Re mfmsr 1 31 83 mfsr 1 31 595 mfsrln 1 31 659 oo mtfsbOx 63 70 Re mtfsb1x 63 38 Re ClfD o ij f f_p_ _ _ _ _ _ _ _ mtfsflx 63 134 Re mtmsr 1 31 146 m...

Страница 408: ...3 31 s A NB 725 M stswx 3 31 s A B 661 1 stwbrx 31 s A B 662 m stwcx 31 s A B 150 stwux 31 s A B 183 stwx 31 s A B 151 sync 31 598 tlbsync 1 5 31 li i i il ii J f g g gi 566 iii 1 1 1F 1F I tw 31 TO...

Страница 409: ...c 19 150 mcrf 19 0 rft 1 19 50 Table A 38 XFX Fom OPCD D spr XO if OPCD D ll CRM I XO 1 OPCD s spr XO I OPCD D tbr XO Specific Instructions Name 0 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 2...

Страница 410: ...A XO Re OPCD D A XO Re Specific Instructions Name 0 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 addx 31 D A B addcx 31 D A B addex 31 D A B addmex 31 D A af l o o negx...

Страница 411: ...faddx 63 D A 21 Re faddsx 59 D A 21 Re fdlvx 63 D A 18 Re tdlvsx 59 D A 18 Re fmaddx 63 D A B C 29 Re fmaddsx 59 D A B C 29 Re fmsubx 63 D A B C 28 Re fmsubsx 59 D A B C 28 Re fmulx 63 D A 25 Re fmul...

Страница 412: ...sh me XO Specific Instructions Name o 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Table A 45 MDS Fonn OPCD s A B mb XO OPCD s A B me XO Specific Instructions Name o 5 6...

Страница 413: ...the 604 A 38 addx addex addex addI addle addle addls addmex addzex andx andex andI andls bx bcx bcctrx bclrx cmp cmpl cmpl cmpll cntlzwx crand erandc UISA Key m Instruction not implemented in the 604...

Страница 414: ...Bit Optional Form ereqv XL ernand XL emor XL eror XL erore XL erxor XL debI x debI x debst x debt x debtst x dcbz x dlvwx XO dlvwux XO eclwx x ecowx x elelo x eqvx x extsbx x extshx x Appendix A Power...

Страница 415: ...tnegx fnmaddx fnmaddsx tnmsubx fnmsubsx A 40 fresx frspx frsqrtex fselx fsubx fsubsx lcbl lsync lbz lbzu lbzux lbzx UISA VEA OEA Supervisor Level 64 Bit Optional Form A A A A x A A A A x x A A A A A...

Страница 416: ...D lfdu D lfdux x lfdx x Ifs D lfsu D lfsux x lfsx x Iha D lhau D lhaux x lhax x lhbrx x lhz D lhzu D lhzux x lhzx x lmw 2 D lswl 2 x lswx 2 x lwbrx x lwz D 1 lwzu D lwzux x lwzx x mcrf XL mcrfs x mcr...

Страница 417: ...mlfsx x mlmsr x mlspr 1 XFX mlsr x mlsrln x mltb XFX mtcrl XFX mtlsbOx x mtlsb1x x mtlslx XFL mtlsllx x mtmsr x mtspr 1 XFX mtsr x mtsrln x mum D mullwx XO nandx x negx XO norx x orx x orcx x Ori D o...

Страница 418: ...mx SC srwx stb stbu stbux stbx stfd sttdu stfdux stfdx stflwx stfs UISA VEA OEA v v v v v v v v v v v Appendix A PowerPC Instruction Set Listings Supervisor Level 64 Bit Optional Form M M M SC x D D x...

Страница 419: ...twcx stwu stwux stwx subfx subfcx subfex subtle subfmex A 44 subfzex sync tw twl UISA VEA OEA Supervisor Level 64 Bit Optional Form v D v x v x v D v x v D v x v x v D v x v x v D v x v x v D v x v x...

Страница 420: ...A VEA OEA Supervisor Level 64 Blt Optional Form 11 l l 1 I I II I 1 Supervisor and user level instruction 2 Load and store string or multiple instruction Appendix A PowerPC Instruction Set Listings A...

Страница 421: ...ion forms of the PowerPC architecture that are not a result of a nonzero reserved field in the instruction encoding Table B 1 Invalid Fonns Excluding Reserved Flelds rA O rAln rAorrB SPRNot Mnemonic 8...

Страница 422: ...ng This table takes into consideration all reserved fields in an instruction that must be zero excluding only those instructions that would become invalid if only bit 31 were set Note that any combina...

Страница 423: ...0 15 20 29 10 15 15 20 20 20 25 bcctr x bcctrl x SC x x mcrf x x x sync x addme o J x subfme o x addze o J x subfze o J x neg o J x mulhw u J x cmpl x x cmp x cmpll x x cmpl x extsb x extsh x cntlzw x...

Страница 424: ...0 29 10 15 15 20 20 20 25 fmuls x fdlv x fdlvs x frsp x fctlw x fctlwz x fcmpu x x fcmpuo x x mffs x mcrfs x x x mtfsfl x x mtfsf x x mtfsbO x mtfsb1 x lcbl x x lsync x x debt x x dcbtst x x dcbz x x...

Страница 425: ...0 20 20 25 mfsrln x x tlble x x mttb x x mttbu x x tlbsync x B 3 Invalid Form with Only Bit 31 Set The following instructions generate invalid instruction forms if only bit 31 is set in the instructio...

Страница 426: ...cla heir bclrl bcctr and bcctrl Specifying a conditional branch instruction with one of these fields results in a invalid instruction form Note that entries with the y bit represent two possible instr...

Страница 427: ...ased exponent s range non negative Big endian A byte ordering method in memory where the address n of a word corresponds to the most significant byte In an addressed memory word the bytes are ordered...

Страница 428: ...icit or implicit leading significand bit is zero E Exception A condition encountered by the processor that requires special Glossary 2 processing Exception handler A software routine that executes whe...

Страница 429: ...nes operations of binary floating point arithmetic and representations of binary floating point numbers Interrupt An asynchronous exception K Kill An operation that causes a cache block to be invalida...

Страница 430: ...Q s Glossary 4 Pipelining A technique that breaks instruction execution into distinct steps so that multiple steps can be performed at the same time Precise exceptions The pipeline can be stopped so t...

Страница 431: ...metic operations when the result cannot be represented accurately in the destination register For example underflow can happen if two floating point fractions are multiplied and the result is a single...

Страница 432: ...al 7 4 8 7 Bits used to configure cache 3 15 Block address translation BAT register initialization 5 13 block address translation flow 5 12 selection of block address translation 5 9 Boundeclly undefi...

Страница 433: ...e 6 11 completion unit 1 9 definition 6 3 Context synchronization 2 24 COP scan interface 7 29 lndex 2 INDEX CR condition register CR logical instructions 2 45 CR description 2 4 CS El signals 7 15 8...

Страница 434: ...6 FP assist exception 4 19 FP unavailable exception 4 4 4 18 instmction address breakpoint exception 4 5 4 20 Index instmction related exceptions 2 25 ISi exception 4 4 machine check 4 3 machine check...

Страница 435: ...h address calculation 2 44 branch instructions A 25 cache management A 26 classes 2 21 condition register logical 2 45 A 25 defined instructions 2 21 eieio 2 49 external control instructions 2 52 A 27...

Страница 436: ...ion instructions A 24 string instructions 2 39 A 24 Index INDEX Load store wiit execution timing 6 39 overview 1 11 Logical addresses translation into physical addresses 5 1 lwarx stwcx M general info...

Страница 437: ...e back stage 6 12 lndex 6 PIR processor identification register 2 8 2 9 PLL configuration 7 31 PMCl and PMC2 registers 2 8 2 13 9 1 9 3 Postdispatch seriali7 ation mode 6 34 Power management nap mode...

Страница 438: ...ister 2 6 Segment registers Index INDEX SR description 2 6 SR manipulation instructions 2 55 A 27 Tbit 8 38 Segmented memory model see Memory management unit SHD signal 7 17 SIA and SDA registers 2 8...

Страница 439: ...ansfers with data delays 8 36 direct store interface load access 8 47 direct store intelface store access 8 48 single beat reads 8 32 single beat reads with data delays 8 34 single beat writes 8 33 si...

Страница 440: ...45 TS signal 7 6 8 12 TSIZO TSIZ2 signals 7 11 8 13 1TO TT4 signals 7 10 8 13 u UISA definition 1 19 registers 2 2 Use ofIBA timing 8 37 User instruction set architecture see UISA Using DBWO timing 8...

Страница 441: ...9 TIME TusUn Time Electronics 1 600 789 TIME West Hiiis Newark 818 888 3718 Woodland Hiiis Hamilton Hallmark 818 594 0404 Richardson Electronics 615 594 5600 COLORADO Colorado Springs Newark 719 592 9...

Страница 442: ...609 424 0100 Fairfield Newark 201 882 0300 Marlton Arrr JN ScJ r NrbJr Electronics 609 596 8000 Future Electronics 609 596 4080 Plnebrook Arrr JN ScJ r NrbJr Electronics 201 227 7880 Wyle Laboratories...

Страница 443: ...Electronics 1 8111 789 TIME Wyle Laboratories 801 974 9953 WASHINGTON Bellewe Almac Electronics Corp 206 643 41992 Newark 206 641 Q800 Richardson Electronics 206 646 7224 Bothell Future Electronics 2...

Страница 444: ...604 293 7650 ONTARIO Toronto 416 497 8181 ONTARIO Ottawa 613 226 3491 QUEBEC Montreal 514 333 3300 INTERNATIONAL AUSTRALIA Melbourne 61 3 887 0711 AUSTRALIA Sydney 61 2 906 3855 BRAZIL Sao Paulo 55 11...

Страница 445: ...Nortec S 46 87051BOO Avnet Nortec OK 45 42 842 000 Avnet Nortec N 47 6684 21o ITT MultikomponentAB 46 8830 020 SINGAPORE Alexan Commercial 63 2 405 e52 GEIC 65 298 7633 P T Ometraco 62 22 63M05 Shapi...

Страница 446: ...9 Fax 800 769 3732 Suite 101 Tel 619 278 4950 Richardson TX 75080 Fax 619 278 0649 131 D Gaither Drive Tel 214 234 8438 Mt Laurel NJ 08054 Fax 214 437 0897 23901 Calabasas Road Tel 609 866 1234 8240 M...

Страница 447: ...875 8219 Dayton Military Tel 602 966 3600 Hartford NW Maryland 446 Windsor Park Drive Fax 602 967 6584 1064 East Main Street Baltimore Mid Atlantic Dayton OH 45459 California Meriden CT 06450 8945 Gu...

Страница 448: ...24 Suite 1024 Tel 414 547 8879 Sacramento Altamonte Springs FL32701 Fax 414 547 6547 Texas 3039 Kilgore Avenue Tel 407 767 8585 100 North Central Expressway Rancho Cordova CA95670 Ste502 Tel 916 635 9...

Страница 449: ...ng Island Austin Toronto Maryland 95 Oser Avenue 8504 Cross Park Drive 4Paget Road Maryland Haupanuge NY 11788 Austin TX 78754 Units 10and 11 2221 Broadbirch Drive Tel 516 273 2695 Tel 512 837 1991 Bu...

Страница 450: ...icroelectronics Department 1045 224 Boulevard J F Kennedy 91105 Corbeil Essonnes CEDEX France Tel 33 HiO 88 5167 Fax 33 1 60 88 4920 Germany ITI Distribution Poslfach 1246 Bahnhofstrasse44 D 71693 Mog...

Страница 451: ...and Bus Interface Unit Operation Exceptions Memory Management Instruction Timing Signal Descriptions System Interface Operation Performance Monitor PowerPC Instruction Set Listings Invalid Instructio...

Страница 452: ...on Exceptions Memory Management Instruction Timing Signal Descriptions System Interface Operation Performance Monitor PowerPC Instruction Set Listing Invalid Instruction Forms Glossary of Terms and Ab...

Страница 453: ...ue revisions to your PowerPC microprocessor documents L _________________ _ ear Here__________________ _J IBM Document Card Company Name_______________________ Your Name_______________________ Title _...

Страница 454: ...NTHE UNITED STATES I II11I11I1I1lI11111II11II11I1I11I1I1111ll1l11I1I1I11I I L ________________ _ _________ _J II BUSINESS REPLY MAIL FIRST CLASS MAIL PERMIT NO 40 ARMONK NY POSTAGE WILL BE PAID BY ADD...

Страница 455: ......

Отзывы:

Нет отзывов

Похожие инструкции для PowerPC 604

Paradyne 3163 Quick Reference preview
3163
Бренд: Paradyne Страницы: 16
2N OfficeRoute Quick Manual preview
OfficeRoute
Бренд: 2N Страницы: 11
2N stargate How To Set preview
stargate
Бренд: 2N Страницы: 5
2N Telekomunikace EasyGate IP Brief Manual preview
EasyGate IP
Бренд: 2N Telekomunikace Страницы: 24
F5 WANJet 500 Quick Start Card preview
WANJet 500
Бренд: F5 Страницы: 2
F5 ARX-500 Hardware Reference Manual preview
ARX-500
Бренд: F5 Страницы: 90
F5 ARX-4000 Parts Replacement Manual preview
ARX-4000
Бренд: F5 Страницы: 47
IEI Technology WSB-H610 User Manual preview
WSB-H610
Бренд: IEI Technology Страницы: 187
IEI Technology TANK-XM811 User Manual preview
TANK-XM811
Бренд: IEI Technology Страницы: 71
T&D RTR-50 User Manual preview
RTR-50
Бренд: T&D Страницы: 122
Garland INT10G10SP2 User Manual preview
INT10G10SP2
Бренд: Garland Страницы: 6
HARTING HAIIC MICA Hardware Development Manual preview
HAIIC MICA
Бренд: HARTING Страницы: 13
Internet Security Systems Proventia A1204 User Manual preview
Proventia A1204
Бренд: Internet Security Systems Страницы: 42
Ubiquiti PBE-5AC-300-ISO Quick Start Manual preview
PBE-5AC-300-ISO
Бренд: Ubiquiti Страницы: 28
pivotal COMMWARE ECHO 5G How To Set Up preview
COMMWARE ECHO 5G
Бренд: pivotal Страницы: 2
IBASE Technology FWA8207 Series User Manual preview
FWA8207 Series
Бренд: IBASE Technology Страницы: 39
Digi ConnectCore 8X Hardware Reference Manual preview
ConnectCore 8X
Бренд: Digi Страницы: 116
AXIOMTEK CEM820 Series User Manual preview
CEM820 Series
Бренд: AXIOMTEK Страницы: 58

Бренды по названию

Популярные бренды

Загрузить еще бренды