Freescale Semiconductor PowerPC e500 Core Скачать руководство пользователя страница 239

Страница: 239 / 548

Execution Timing

PowerPC e500 Core Family Reference Manual, Rev. 1

Freescale Semiconductor

4-37

lbz

3:1 —

lbzu

3:1

—

lbzux

3:1

—

lbzx

3:1

—

lha

3:1

—

lhau

3:1

—

lhaux

3:1

—

lhax

3:1

—

lhbrx

3:1

—

lhz

3:1

—

lhzu

3:1

—

lhzux

3:1

—

lhzx

3:1

—

lmw

2 + n

—

lwarx

Presync

lwbrx

3:1

—

lwz

3:1

—

lwzu

3:1

—

lwzux

3:1

—

lwzx

3:1

—

mbar

3:1

Store

serialized

msync

Latency depends on bus response time.

Store and postsync serialized.

stb

3:1

Store

stbu

3:1

Store

stbux

3:1

Store

stbx

3:1

Store

sth

3:1

Store

sthbrx

Store

sthu

3:1

Store

sthux

3:1

Store

sthx

3:1

Store

stmw

3 + n

Store

stw

3:1

Store

stwbrx

3:1

Store

Table 4-7. LSU Instruction Latencies (continued)

Mnemonic

Cycles (Latency:Throughput)

Serialization

Содержание PowerPC e500 Core

Страница 1: ...PowerPC e500 Core Family Reference Manual Supports e500v1 e500v2 E500CORERM Rev 1 4 2005...

Страница 2: ...for surgical implant into the body or other applications intended to support or sustain life or for any other application in which the failure of the Freescale Semiconductor product could create a si...

Страница 3: ...Debug Support 8 Part II e500 Core Complex II Timer Facilities 9 Auxiliary Processing Units APUs 10 L1 Caches 11 Memory Management Units 12 Core Complex Bus CCB 13 Appendix A Programming Examples A Ap...

Страница 4: ...r 8 Debug Support II Part II e500 Core Complex 9 Timer Facilities 10 Auxiliary Processing Units APUs 11 L1 Caches 12 Memory Management Units 13 Core Complex Bus CCB A Appendix A Programming Examples B...

Страница 5: ...Processor and System Version Numbers 1 5 1 3 Features 1 5 1 3 1 e500v2 Differences 1 11 1 4 Instruction Set 1 12 1 5 Instruction Flow 1 14 1 5 1 Initial Instruction Fetch 1 14 1 5 2 Branch Detection...

Страница 6: ...13 1 Instruction Set Compatibility 1 32 1 13 1 1 User Instruction Set 1 32 1 13 1 2 Supervisor Instruction Set 1 32 1 13 2 Memory Subsystem 1 33 1 13 3 Exception Handling 1 33 1 13 4 Memory Managemen...

Страница 7: ...2 1 Machine Check Save Restore Register 0 MCSRR0 2 22 2 7 2 2 Machine Check Save Restore Register 1 MCSRR1 2 22 2 7 2 3 Machine Check Address Register MCAR 2 22 2 7 2 4 Machine Check Syndrome Register...

Страница 8: ...ug Status Register DBSR 2 47 2 13 3 Instruction Address Compare Registers IAC1 IAC4 2 48 2 13 4 Data Address Compare Registers DAC1 DAC2 2 48 2 14 SPE and SPFP APU Registers 2 49 2 14 1 Signal Process...

Страница 9: ...Integer Rotate and Shift Instructions 3 16 3 3 1 2 Load and Store Instructions 3 17 3 3 1 2 1 Self Modifying Code 3 17 3 3 1 2 2 Integer Load and Store Address Generation 3 18 3 3 1 2 3 Integer Load...

Страница 10: ...cquisition and Import Barriers 3 45 3 5 1 1 Acquire Lock and Import Shared Memory 3 45 3 5 1 2 Obtain Pointer and Import Shared Memory 3 45 3 5 1 3 Lock Release and Export Barriers 3 46 3 5 1 3 1 Expo...

Страница 11: ...3 5 Memory Synchronization Timing Considerations 4 17 4 3 5 1 msync Instruction Timing Considerations 4 17 4 3 5 2 mbar Instruction Timing Considerations 4 17 4 4 Execution 4 18 4 4 1 Branch Unit Exe...

Страница 12: ...ons 4 47 4 7 6 1 SU Considerations 4 47 4 7 6 2 MU Considerations 4 48 4 7 6 3 LSU Considerations 4 48 4 7 6 3 1 Load Store Interaction 4 48 4 7 6 3 2 Misalignment Effects 4 49 4 7 6 3 3 Load Miss Pip...

Страница 13: ...ance Monitor Interrupt 5 33 5 9 Partially Executed Instructions 5 33 5 10 Interrupt Ordering and Masking 5 35 5 10 1 Guidelines for System Software 5 36 5 10 2 Interrupt Order 5 37 5 11 Exception Prio...

Страница 14: ...7 Event Selection 7 12 Chapter 8 Debug Support 8 1 Overview 8 1 8 2 Programming Model 8 1 8 2 1 Register Set 8 1 8 2 2 Instruction Set 8 2 8 2 3 Debug Interrupt Model 8 2 8 2 4 Deviations from the Boo...

Страница 15: ...ogramming Model 10 2 10 2 1 1 BTB Locking APU Instructions 10 2 10 2 1 2 BTB Locking APU Registers 10 3 10 3 Alternate Time Base APU 10 3 10 3 1 Programming Model 10 3 10 4 Double Precision Floating P...

Страница 16: ...struction Cache Coherency Model 11 11 11 3 3 Snoop Signaling 11 12 11 3 4 WIMGE Settings and Effect on L1 Caches 11 13 11 3 4 1 Write Back Stores 11 13 11 3 4 2 Write Through Stores 11 13 11 3 4 3 Cac...

Страница 17: ...a Cache Tag Parity Error 11 27 11 7 2 L2 Locking 11 27 11 7 2 1 L2 Unlocking 11 28 11 7 2 2 L1 Overlock 11 28 Chapter 12 Memory Management Units 12 1 e500 MMU Overview 12 1 12 1 1 MMU Features 12 1 1...

Страница 18: ...2 4 4 3 TLB Invalidate Broadcast Enabling 12 22 12 4 5 TLB Synchronize tlbsync Instruction 12 22 12 5 TLB Entry Maintenance Details 12 22 12 5 1 Automatic Updates TLB Miss Exceptions 12 23 12 5 2 TLB...

Страница 19: ...r 32 Bit Book E B 1 64 Bit Specific Book E Instructions B 1 B 2 Registers on 32 Bit Book E Implementations B 2 B 3 Addressing on 32 Bit Book E Implementations B 2 B 4 TLB Fields on 32 bit Book E Imple...

Страница 20: ...nics that Incorporate CR Conditions Listings C 17 C 5 Compare Word Simplified Mnemonics C 20 C 6 Condition Register Logical Simplified Mnemonics C 20 C 7 Trap Instructions Simplified Mnemonics C 21 C...

Страница 21: ...ter Upper ATBU 2 17 2 9 Interrupt Vector Offset Registers IVORs 2 19 2 10 Exception Syndrome Register ESR 2 20 2 11 Machine Check Save Restore Register 0 MCSRR0 2 22 2 12 Machine Check Save Restore Re...

Страница 22: ...MC3 User Performance Monitor Counter Registers UPMC0 UPMC3 2 57 3 1 Register Indirect with Immediate Index Addressing for Integer Loads Stores 3 18 3 2 Register Indirect with Index Addressing for Inte...

Страница 23: ...4 PLRU Replacement Algorithm 11 26 12 1 Effective to Real Address Translation Flow e500v1 12 4 12 2 Effective to Real Address Translation Flow e500v2 12 5 12 3 Virtual Address and TLB Entry Compare Pr...

Страница 24: ...PowerPC e500 Core Family Reference Manual Rev 1 xxiv Freescale Semiconductor Figures Figure Number Title Page Number...

Страница 25: ...2 9 ESR Field Descriptions 2 21 2 10 MCSR Field Descriptions 2 23 2 11 BBEAR Field Descriptions 2 25 2 12 BBTAR Field Descriptions 2 25 2 13 BUCSR Field Descriptions 2 26 2 14 HID0 Field Descriptions...

Страница 26: ...on Requirements 3 9 3 6 Integer Arithmetic Instructions 3 14 3 7 Integer 32 Bit Compare Instructions L 0 3 15 3 8 Integer Logical Instructions 3 15 3 9 Integer Rotate Instructions 3 16 3 10 Integer Sh...

Страница 27: ...es 4 31 4 4 System Operation Instruction Execution Latencies 4 31 4 5 Condition Register Logical Execution Latencies 4 33 4 6 SU and MU PowerPC Instruction Execution Latencies 4 33 4 7 LSU Instruction...

Страница 28: ...ettings 5 31 5 30 Embedded Floating Point Data Interrupt Register Settings 5 32 5 31 Embedded Floating Point Round Interrupt Register Settings 5 33 5 32 Operations to Avoid 5 36 6 1 Power Management S...

Страница 29: ...12 10 MAS2 Field Descriptions EPN and Page Attributes 12 28 12 11 MAS3 Field Descriptions RPN and Access Control 12 29 12 12 MAS4 Field Descriptions Hardware Replacement Assist Configuration 12 30 12...

Страница 30: ...Simplified Mnemonics for bcl and bcla with Comparison Conditions and LR Updating C 18 C 22 Simplified Mnemonics for bclrl and bcctrl with Comparison Conditions and LR Updating C 19 C 23 Word Compare S...

Страница 31: ...llows The Book E architecture Book E defines a set of user level instructions and registers that are drawn from the user instruction set architecture UISA portion of the AIM definition PowerPC archite...

Страница 32: ...issues executed and completed and how instruction results are presented to the processor and memory system Tables are provided that indicate latency and throughput for each of the instructions support...

Страница 33: ...ics for PowerPC Instructions provides a set of simplified mnemonic examples and symbols Appendix D Opcode Listings lists opcodes by mnemonic and by opcode It includes an alphabetical listing that incl...

Страница 34: ...Hardware specifications Hardware specifications provide specific data regarding bus timing signal behavior and AC DC and thermal characteristics as well as other design considerations Product briefs E...

Страница 35: ...e NOT logical operator AND logical operator OR logical operator Indicates reserved bits or bit fields in a register Although these bits can be written to as ones or zeros they are always read as zeros...

Страница 36: ...PowerPC e500 Core Family Reference Manual Rev 1 xxxvi Freescale Semiconductor...

Страница 37: ...he e500 core s operations performance as defined by instructions and how it reports the results of instruction execution It gives detailed descriptions of how the core execution units work and how the...

Страница 38: ...PowerPC e500 Core Family Reference Manual Rev 1 Part I 2 Freescale Semiconductor...

Страница 39: ...summary of the Book E architecture compatibility and migration from the original version of the PowerPC architecture as it is defined by Apple IBM and Motorola referred to as the AIM version of the Po...

Страница 40: ...Issue Queue BIQ TLB0 TLB Array TLB1 Register Unified L1 Data MMU 64 Entry D L1TLB4K 4 Entry D L1VSP Two Instruction Dispatch 1 BIQ 2 GIQ Core Interface Unit L1 Instruction MMU 64 Entry I L1TLB4K 4 En...

Страница 41: ...unctionality is implemented in all PowerQUICC III devices However these instructions will not be supported in devices subsequent to PowerQUICC III Freescale Semiconductor strongly recommends that use...

Страница 42: ...support 4 Kbyte pages These arrays are maintained entirely by the hardware with a true least recently used LRU algorithm The second level MMU contains a 16 entry fully associative unified instruction...

Страница 43: ...r facility provides the ability to monitor and count predefined events such as processor clocks misses in the instruction cache or data cache types of instructions decoded or mispredicted branches The...

Страница 44: ...d to 64 bits to support 64 bit load store and merge operations Note that the upper 32 bits are affected only by 64 bit instructions A 64 bit accumulator ACC has been added The signal processing and em...

Страница 45: ...of interlocks Decodes as many as two instructions per cycle Decode serialization control Register dependency resolution and renaming Branch prediction unit BPU Dynamic branch prediction using a 512 e...

Страница 46: ...divide instructions If rA or rB is zero floating point divide instructions take 4 cycles all others take 29 Note that although most divide instructions take more than 4 cycles to execute the MU allow...

Страница 47: ...in some implementations Bus support for hardware enforced coherency bus snooping Core complex bus CCB internal bus High speed on chip local bus with data tagging 32 bit address bus Address protocol w...

Страница 48: ...for the e500v1 and 4 Kbyte 4 Gbyte pages for the e500v2 and fixed size 4 Kbyte pages Data L1 MMU 4 entry fully associative TLB array for variable sized pages 64 entry 4 way set associative TLB for 4...

Страница 49: ...itional features not supported by the e500v1 The e500v2 uses 36 bit physical addressing which is supported by the following MMU assist register 7 MAS7 HID0 EN_MAS7_UPDATE Programmable jumper options t...

Страница 50: ...e500 supports the following implementation specific instructions Integer select APU This APU consists of the Integer Select instruction isel which functions as an if then else statement that selects...

Страница 51: ...ed Integer efscfsi efdcfsi evfscfsi rD rB Convert Floating Point from Unsigned Fraction efscfuf efdcfuf evfscfuf rD rB Convert Floating Point from Unsigned Integer efscfui efdcfui evfscfui rD rB Conve...

Страница 52: ...ructions using both the lower and upper halves of the 64 bit GPRs The parallel execution units allow multiple instructions to execute in parallel and out of order For example a low latency addition in...

Страница 53: ...t streams are allowed to execute and proceed through the completion queue although they can complete only after the branch prediction is resolved as correct and after the branch instruction itself com...

Страница 54: ...tion is in the on chip instruction cache or an L2 cache if implemented Those factors increase when it is necessary to fetch instructions from system memory and include the processor to bus clock ratio...

Страница 55: ...that are assigned a space in the CQ but not in an issue queue Dispatch is treated as an event at the end of the decode stage The issue stage reads source operands from rename registers and register fi...

Страница 56: ...ng an exception status or a mispredicted branch all following instructions are cancelled their execution results in rename registers are discarded and the correct instruction stream is fetched The com...

Страница 57: ...RR0 Critical SRR 0 1 Processor version spr 528 IVOR32 1 Interrupt vector offset registers 32 35 spr 287 PVR spr 59 CSRR1 spr 529 IVOR33 1 Timer Decrementer Registers spr 570 MCSRR0 1 Machine check SRR...

Страница 58: ...etting L1CSR0 CFI the entire instruction cache can be invalidated by setting L1CSR1 ICFI 1 8 Interrupts and Exception Handling The e500 core supports an extended exception handling model with nested i...

Страница 59: ...zing mechanism the address in the appropriate save restore register is the address of the interrupt forcing instruction If the interrupt was not caused by either of those mechanisms the address in the...

Страница 60: ...ase the latency is indeterminate The minimum latency is 3 core clocks and the maximum is 8 not including the 2 bus clock cycles required to synchronize the interrupt signal from the pad When an interr...

Страница 61: ...ation on machine check interrupts and restores machine state after an rfmci instruction is executed ESR Exception syndrome register Provides a syndrome to differentiate between the different kinds of...

Страница 62: ...ough required W Caching inhibited I Memory coherency required M Guarded G Endianness E User definable U0 U3 a 4 bit implementation specific field The core complex employs a two level memory management...

Страница 63: ...2 TLB array TLB1 that supports all nine variable page sizes TLB array TLB0 that supports only 4 Kbyte pages as follows e500v1 256 entry 2 way set associative TLB array e500v2 512 entry 4 way set assoc...

Страница 64: ...t virtual address Figure 1 9 shows the translation flow for the e500v1 core Figure 1 9 Effective to Real Address Translation Flow Effective Page Number Byte Address Real Page Number Byte Address 32 bi...

Страница 65: ...rts using a true LRU algorithm 1 9 2 MMU Assist Registers MAS0 MAS4 and MAS6 MAS7 Book E defines SPR numbers for the MMU assist registers which are used to hold values either read from or to be writte...

Страница 66: ...MMU The lookup is specified by the instruction encoding and specific search fields in MAS6 The values placed in the MAS registers may differ depending on a successful or unsuccessful search For TLB m...

Страница 67: ...ionality The processor revokes reservations during a context switch so the programmer must reacquire the reservation after a context switch occurs 1 10 2 Memory Access Ordering The core complex suppor...

Страница 68: ...wo different sources simultaneously and all three data buses may be operated concurrently The address in bus supports snooping for external management of the L1 caches and TLBs by other bus masters Th...

Страница 69: ...ollowing conditions are met A counter is in the overflow state The counter s overflow signaling is enabled Overflow exception generation is enabled in PMGC0 MSR EE is set 1 12 3 Local Control Register...

Страница 70: ...1 Instruction Set Compatibility The following sections generally describe the user and supervisor instruction sets 1 13 1 1 User Instruction Set The e500 core executes legacy user mode binaries and ob...

Страница 71: ...efined in the OEA to provide compatibility Unlike the AIM version of the PowerPC architecture Book E does not define a reset vector execution begins at a fixed virtual address 0xFFFF_FFFC Some Book E...

Страница 72: ...otection mechanisms Unlike the AIM version of the PowerPC core as soon as instruction fetching begins the e500 core is in virtual mode with a hardware initialized TLB entry EIS defined aspects of the...

Страница 73: ...mplemented on the device 2 1 Overview Although this chapter organizes registers according to their functionality they can be differentiated according to how they are accessed as follows General purpos...

Страница 74: ...registers Book E defined registers that are accessed as part of instruction execution These include the following Registers used for integer operations General purpose registers GPRs Book E defines a...

Страница 75: ...tical SRR 0 1 Processor version spr 528 IVOR32 1 Interrupt vector offset registers 32 35 spr 287 PVR spr 59 CSRR1 spr 529 IVOR33 1 Timer Decrementer Registers spr 570 MCSRR0 1 Machine check SRR 0 1 sp...

Страница 76: ...2 8 Software Use SPRs SPRG0 SPRG7 and USPRG0 Section 2 9 Branch Target Buffer BTB Registers Section 2 10 Hardware Implementation Dependent Registers Section 2 11 L1 Cache Configuration Registers Secti...

Страница 77: ...ause a principal goal of the Book E architecture is to offer flexibility among embedded processors and families of embedded processors some resources are either defined as optional or are defined in a...

Страница 78: ...SR Exception syndrome register 62 00001 11110 Read Write Yes 2 7 1 6 2 20 IAC1 Instruction address compare 1 312 01001 11000 Read Write Yes 2 13 3 2 48 IAC2 Instruction address compare 2 313 01001 110...

Страница 79: ...18 TBL Time base lower 268 01000 01100 Read only No 2 6 3 2 16 284 01000 11100 Write only Yes 2 6 3 2 16 TBU Time base upper 269 01000 01101 Read only No 2 6 3 2 16 285 01000 11101 Write only Yes 2 6...

Страница 80: ...register 1 516 Read only No 2 11 4 2 35 L1CSR0 L1 cache control and status register 0 1 1010 Read Write Yes 2 11 1 2 31 L1CSR1 L1 cache control and status register 1 1 1011 Read Write Yes 2 11 2 2 33...

Страница 81: ...ation of an instruction considered as a whole not on intermediate results For example the Subtract from Carrying instruction subfc the result of which is specified as the sum of three values sets bits...

Страница 82: ...e 2 2 defines the state of the processor that is enabling and disabling of interrupts and debugging exceptions enabling and disabling of address translation for instruction and data memory accesses en...

Страница 83: ...uctions that access the upper half of the 64 bit GPRs 39 44 Reserved should be cleared 1 45 WE Wait state enable On the e500 this allows the core complex to signal a request for power management accor...

Страница 84: ...enabled 54 DE Debug interrupt enable 0 Debug interrupts are disabled 1 Debug interrupts are enabled if DBCR0 IDM 1 For the e500 see the description of the DBSR UDE in Section 2 13 2 Debug Status Regi...

Страница 85: ...Revision W Reset e500v1 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 Reset e500v2 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 Figure 2 3 Processor Version Regis...

Страница 86: ...e software timer The time base for the TB and DEC is selected by the time base enable TBEN and select time base clock SEL_TBCLK bits in HID0 as follows If HID0 TBEN 1 and HID0 SEL_TBCLK 0 the time bas...

Страница 87: ...FIE ARE WPEXT FPEXT W Reset All zeros Figure 2 6 Timer Control Register TCR Table 2 5 TCR Implementation Specific Field Descriptions Bits Name Description 32 33 WP Watchdog timer period When concaten...

Страница 88: ...one of the following occurs DEC is altered by software in the interim The TB update frequency changes DEC is typically used as a general purpose software timer The decrementer auto reload register is...

Страница 89: ...is accessible in both user and supervisor mode Table 2 7 describes the ATBU fields 2 7 Interrupt Registers Section 2 7 1 Interrupt Registers Defined by Book E and Section 2 7 2 e500 Specific Interrupt...

Страница 90: ...hen rfi executes SRR1 contents are placed into MSR SRR1 bits that correspond to reserved MSR bits are also reserved These registers are not affected by rfci or rfmci Reserved MSR bits may be altered b...

Страница 91: ...rrupt type Table 2 8 shows the IVORs implemented on the e500 IVOR0 IVOR15 are defined by the architecture Note that the e500 does not implement IVOR7 and IVOR9 In addition IVOR32 IVOR35 SPR 528 531 ar...

Страница 92: ...se the ESR for machine check interrupts but instead uses the machine check syndrome register MCSR described in Section 2 7 2 4 Machine Check Syndrome Register MCSR The ESR is defined in Book E but dif...

Страница 93: ...39 Not supported on the e500 Defined by Book E as FP floating point operations On the e500 this bit is reserved and permanently cleared indicating that the e500 does not implement a Book E FPU Setting...

Страница 94: ...achine check interrupt is taken MSR contents are placed into MCSRR1 shown in Figure 2 12 When rfmci executes MCSRR1 contents are restored to MSR MCSRR1 bits that correspond to reserved MSR bits are al...

Страница 95: ...corresponding module The MCSR is shown in Figure 2 14 Table 2 10 describes the MCSR fields SPR 572 Access Supervisor only 32 33 34 35 36 39 R MCP ICPERR DCP_PERR DCPERR W Reset All zeros 40 47 R W Re...

Страница 96: ...ster can be accessed in supervisor or user mode 2 9 Branch Target Buffer BTB Registers SPRs are defined in the core complex for enabling the locking and unlocking of entries in the BTB These are calle...

Страница 97: ...ve entry address bits 0 29 62 63 IAB 0 1 Instruction after branch with BBTAR 62 3 bit pointer that points to the instruction in the cache block after the branch If the branch is the last instruction i...

Страница 98: ...FI flash clears the valid bit of all entries in the branch buffer clearing occurs independently from the value of the enable bit BPEN BBFI is always read as 0 55 BBLO Branch buffer lock overflow statu...

Страница 99: ...onization as described in Section 2 16 Synchronization Requirements for SPRs Table 2 14 describes the HID0 fields SPR 1008 Access Supervisor only 32 33 39 40 41 42 43 47 R EMCP DOZE NAP SLEEP W Reset...

Страница 100: ...tlbre and tlbsx 0 MAS7 is not updated by a tlbre or tlbsx 1 MAS7 is updated by a tlbre or tlbsx 57 DCFA Data cache flush assist e500v2 only Force data cache to ignore invalid sets on miss replacement...

Страница 101: ...r to the Register Summary chapter in the device s reference manual HID1 is used for bus configuration and control Writing to HID1 requires synchronization as described in Section 2 16 Synchronization...

Страница 102: ...n is taken immediately and processing does not continue with potentially bad data However setting RFXE when a peripheral block is configured to also signal an interrupt for a core_fault_in case result...

Страница 103: ...dcbf mbar msync tlbivax tlbsync icbi based on ABE On some implementations ABE must be set to allow management of external L2 caches 0 Address broadcasting disabled 1 Address broadcasting enabled 52 R...

Страница 104: ...tered a dcbi snoop that invalidated a locked line 1 The cache has encountered a dcbi snoop that invalidated a locked line 53 CUL Data Cache unable to lock Sticky bit set by hardware and cleared by wri...

Страница 105: ...must also be enabled ICPE 1 when this bit is set Note that if the programmer attempts to set L1CSR1 ICPI using mtspr without setting L1CSR1 ICPE L1CSR1 ICPI will not be set enforced by hardware 49 51...

Страница 106: ...cessed or updated 1 Enables instruction cache operation SPR 515 Access Supervisor read only 32 33 34 38 39 40 41 42 43 44 45 49 50 51 52 53 55 56 63 R CARCH CBSIZE CREPL CLA CPA CNWAY CSIZE W Reset 0...

Страница 107: ...ZE Cache size 0x20 indicates 32 Kbytes SPR 516 Access Supervisor read only 32 38 39 40 41 42 43 44 45 52 53 63 R ICBSIZE ICREPL ICLA ICPA ICNWAY ICSIZE W Reset 0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 1 1...

Страница 108: ...and L2 MMUs Writing to MMUCSR0 requires synchronization as described in Section 2 16 Synchronization Requirements for SPRs Table 2 20 describes the MMUCSR0 fields SPR SPR SPR 48 PID0 PID in Book E 633...

Страница 109: ...ers A 4 bit field that indicates the number of PID registers provided by the processor The e500 implements three PIDs 53 57 PIDSIZE PID register size The 5 bit value of PIDSIZE is one less than the nu...

Страница 110: ...0 0 0 0 0 Figure 2 27 TLB Configuration Register 0 TLB0CFG Table 2 22 TLB0CFG Field Descriptions Bits Name Description 32 39 ASSOC Associativity of TLB0 0x02 Indicates associativity is 2 way set assoc...

Страница 111: ...3 R ASSOC MINSIZE MAXSIZE IPROT AVAIL NENTRY W Reset e500v1 0 0 0 1 0 0 0 0 0 0 0 1 1 0 0 1 1 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 Reset e500v2 0 0 0 1 0 0 0 0 0 0 0 1 1 0 1 1 1 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0...

Страница 112: ...LB as follows When TLBSEL 00 TLB0 selected bits 46 47 are used and bits 44 45 should be cleared This field selects between way 0 1 2 or 3 of TLB0 EA bits 45 51 from MAS2 EPN are used to index into the...

Страница 113: ...rays that support invalidate protection are denoted as such in the TLB configuration registers 0 Entry is not protected from invalidation 1 Entry is protected from invalidation 34 39 Reserved should b...

Страница 114: ...Caching inhibited 0 Accesses to this page are considered cacheable 1 The page is considered caching inhibited All loads and stores to the page bypass the caches and are performed directly to main mem...

Страница 115: ...s Name Description 32 51 RPN Real page number Depending on page size only the bits associated with a page boundary are valid Bits that represent offsets within a page are ignored and should be zero No...

Страница 116: ...on The e500 implementation defines bits 44 45 as reserved and bits 46 47 as follows 00 PID0 01 PID1 10 PID2 11 TIDZ 0x00 all zeros 48 51 Reserved should be cleared 52 55 TSIZED Default TSIZE value Spe...

Страница 117: ...or These registers are intended for use by special debug tools and debug software and not by general application or operating system code Table 2 29 MAS6 Field Descriptions Bits Name Description 32 39...

Страница 118: ...uires synchronization as described in Section 2 16 Synchronization Requirements for SPRs Table 2 32 describes the DBCR1 fields Table 2 31 DBCR0 Field Descriptions Bits Name Description 34 35 RST Reset...

Страница 119: ...le 2 34 DBSR UDE The MRR field is affected by the e500 definition of the HRESET signal as defined in Table 2 34 SPR 310 Access Supervisor only 32 33 34 35 36 37 38 39 40 41 42 63 R DAC1US DAC1ER DAC2U...

Страница 120: ...s Compare Registers DAC1 DAC2 The e500 implements the DAC1 and DAC2 as they are defined by the Book E architecture A debug event may be enabled to occur upon loads stores or cache operations to an add...

Страница 121: ...PowerQUICC III Freescale Semiconductor strongly recommends that use of these instructions be confined to libraries and device drivers Customer software that uses SPE or embedded floating point APU in...

Страница 122: ...undefined if the processor takes a floating point exception due to input error floating point overflow or floating point underflow 35 FXH Embedded floating point sticky bit high Floating bit from the...

Страница 123: ...dividend and divisor are zero 53 FDBZ Embedded floating point divide by zero error Set if the dividend is non zero and the divisor is zero 54 FUNF Embedded floating point underflow error 55 FOVF Embed...

Страница 124: ...urces used exclusively by the performance monitor PMRs are similar to the SPRs defined in the Book E architecture and are accessed by mtpmr and mfpmr which are also defined by the EIS Table 2 36 lists...

Страница 125: ...0 2 15 4 2 56 UPMLCa1 User performance monitor local control a1 129 00100 00001 UPMLCa2 User performance monitor local control a2 130 00100 00010 UPMLCa3 User performance monitor local control a3 131...

Страница 126: ...nt occurs when the selected bit changes from 0 to 1 e500v2 only 00 TB 63 TBL 31 01 TB 55 TBL 23 10 TB 51 TBL 19 11 TB 47 TBL 15 Time base transition events can be used to periodically collect informat...

Страница 127: ...ter 0 The PMC is incremented if permitted by other PM control bits 1 The PMC is not incremented 33 FCS Freeze counter in supervisor state 0 The PMC is incremented if permitted by other PM control bits...

Страница 128: ...le 2 40 PMLCb0 PMLCb3 Field Descriptions Bits Name Description 32 52 Reserved should be cleared 53 55 THRESHMUL Threshold multiple 000 Threshold field is multiplied by 1 PMLCbn THRESHOLD 1 001 Thresho...

Страница 129: ...le MSR EE is zero is not taken until MSR EE is set Setting PMGC0 FCECE forces counters to stop counting when an enabled condition or event occurs Software is expected to use mtpmr to explicitly set PM...

Страница 130: ...for these SPRs there are no synchronization requirements for accessing SPRs beyond those stated in Book E Table 2 42 Synchronization Requirements for SPRs Registers Instruction Instruction Required B...

Страница 131: ...pter 10 Auxiliary Processing Units APUs Specific information about how these instructions are executed is provided in Chapter 4 Execution Timing 3 1 Operand Conventions This section describes operand...

Страница 132: ...te that lmw stmw lwarx and stwcx instructions that are not word aligned cause an alignment exception 3 1 3 e500 Floating Point Implementation The e500 does not implement the floating point instruction...

Страница 133: ...tions Freescale will also provide libraries to support next generation PowerQUICC devices 3 1 4 Unsupported Book E Instructions Because the e500 core complex uses a 32 bit Book E core all of the instr...

Страница 134: ...d CR fmul s Floating Negative Absolute Value and record CR fnabs Floating Negate and record CR fneg Floating Negative Multiply Add Single and record CR fnmadd s Floating Negative Multiply Subtract Sin...

Страница 135: ...ons Memory synchronization instructions These instructions are used for memory synchronizing See Section 3 3 1 6 Memory Synchronization Instructions Memory control instructions These instructions prov...

Страница 136: ...s Preserved instructions Reserved illegal or no op instructions These classes are defined in the Instruction Model chapter of the EREF The class is determined by examining the primary opcode and any e...

Страница 137: ...context altering instruction should be interpreted as meaning the context altering instruction itself The synchronizing instruction before the context altering instruction ensures that all instructio...

Страница 138: ...efore may not be context synchronizing mtmsr DS None CSI 1 mtmsr WE msync isync 3 3 See Section 6 4 1 Software Considerations for Power Management mtspr DAC1 DAC2 4 4 Synchronization requirements for...

Страница 139: ...synchronizing instruction that is sc isync rfci or rfi Table 3 5 Instruction Fetch and or Execution Synchronization Requirements Context Altering Instruction or Event Required Before Required After N...

Страница 140: ...may not be context synchronizing 4 CSI indicates any context synchronizing instruction that is sc isync rfci rfmci or rfi 5 Synchronization requirements for changing the wait state enable are implemen...

Страница 141: ...l address space and memory protection in which they were initiated 3 If the operation directly causes an interrupt for example sc directly causes a system call interrupt or is an interrupt the operati...

Страница 142: ...rrupt An attempt by an application program to access a privileged SPR privileged instruction exception type program interrupt An attempt by an application program to access an SPR that does not exist...

Страница 143: ...errupt conditions in detail 3 3 Instruction Set Overview This section provides a overview of the PowerPC instructions implemented in the e500 and highlights any special information with respect to how...

Страница 144: ...tion of a 32 bit result only if the instruction s OE bit is set Table 3 6 Integer Arithmetic Instructions Name Mnemonic Syntax Add add add addo addo rD rA rB Add Carrying addc addc addco addco rD rA r...

Страница 145: ...ctions The logical instructions shown in Table 3 8 perform bit parallel operations on the specified operands Logical instructions with the CR updating enabled uses dot suffix and instructions andi and...

Страница 146: ...s are obtained by specifying masks and shift values for certain rotate instructions Simplified mnemonics shown in Appendix C Simplified Mnemonics for PowerPC Instructions are provided to simplify codi...

Страница 147: ...t cross a double word boundary degrade performance Although many misaligned memory accesses are supported in hardware the frequent use of them is discouraged because they can compromise the overall pe...

Страница 148: ...hes 3 3 1 2 2 Integer Load and Store Address Generation Integer load and store operations generate effective addresses using register indirect with immediate index mode register indirect with index mo...

Страница 149: ...instruction descriptions as rA 0 Figure 3 2 shows how an effective address is generated using this addressing mode Figure 3 2 Register Indirect with Index Addressing for Integer Loads Stores Register...

Страница 150: ...structions Name Mnemonic Syntax Load Byte and Zero lbz rD d rA Load Byte and Zero Indexed lbzx rD rA rB Load Byte and Zero with Update lbzu rD d rA Load Byte and Zero with Update Indexed lbzux rD rA r...

Страница 151: ...e page is marked cacheable write through WIM 10x but as with other memory accesses data storage interrupts can result for other reasons such as protection violations or page faults 3 3 1 2 4 Integer S...

Страница 152: ...structions is interrupted it may be restarted requiring multiple memory accesses The Book E architecture defines the Load Multiple Word lmw instruction with rA in the range of registers to be loaded a...

Страница 153: ...O operand is ignored for branch prediction The e500 instead implements dynamic branch prediction as part of the branch table buffer BTB described in Section 4 4 1 Branch Unit Execution The encodings f...

Страница 154: ...simplified mnemonics and symbols is provided for the most frequently used forms of branch conditional compare trap rotate and shift and certain other instructions see Appendix C Simplified Mnemonics...

Страница 155: ...e trap instructions shown in Table 3 19 test for a specified set of conditions If any of the conditions tested by a trap instruction are met the system trap type program interrupt is taken For more in...

Страница 156: ...izes the instructions for reading from or writing to the CR Implementation Note The Book E architecture states that the Move to Condition Register Fields mtcrf instruction can perform more slowly when...

Страница 157: ...ompare 1 312 01001 11000 Read Write Yes 2 13 3 2 48 IAC2 Instruction address compare 2 313 01001 11001 Read Write Yes 2 13 3 2 48 IVOR0 Critical input 400 01100 10000 Read Write Yes 2 7 1 5 2 19 IVOR1...

Страница 158: ...ower 268 01000 01100 Read only No 2 6 3 2 16 284 01000 11100 Write only Yes 2 6 3 2 16 TBU Time base upper 269 01000 01101 Read only No 2 6 3 2 16 285 01000 11101 Write only Yes 2 6 3 2 16 TCR Timer c...

Страница 159: ...4 2 35 L1CSR0 L1 cache control and status register 0 1010 Read Write Yes 2 11 1 2 31 L1CSR1 L1 cache control and status register 1 1011 Read Write Yes 2 11 2 2 33 MAS0 MMU assist register 0 624 Read W...

Страница 160: ...tests selected CRn EQ and even if the branch target is the next sequential instruction to be executed Load Word and Reserve Indexed lwarx rD rA rB lwarx with stwcx can emulate semaphore operations suc...

Страница 161: ...el conditions Frequent use of msync degrades performance System designs with an external cache should take care to recognize the hardware signaling caused by an MSYNC bus operation and perform the app...

Страница 162: ...altered only by stores that are in the same set for both cases described above For the second use mbar MO 1 can be thought of as placing a barrier into the stream of memory accesses issued by a core s...

Страница 163: ...x The stwcx is a store to a word aligned location that is conditioned on the existence of the reservation created by the lwarx and on whether both instructions specify the same location To emulate an...

Страница 164: ...and swap primitive like that provided by the IBM System 370 compare and swap instruction which checks only that the old and current values of the word being tested are equal with the result that progr...

Страница 165: ...ntended to be paired must be scrupulously avoided For example there must not be a context switch in which the processor holds a reservation on behalf of the old context and the new context resumes aft...

Страница 166: ...context switches It may also include implementation dependent causes of reservation loss An implementation may make a forward progress guarantee defining the conditions under which the system as a who...

Страница 167: ...ons The instructions listed in Table 3 26 help user level programs manage on chip caches if they are implemented See Chapter 11 L1 Caches for more information about cache topics The following sections...

Страница 168: ...formed and the cache block is marked modified However if the replacement block is marked modified the contents are written back to memory first dcbz takes an alignment interrupt if the cache is locked...

Страница 169: ...in the L2 cache In this case dcbtst reloads the L1 data cache with the state found in the L2 cache Again if the block was in the shared state in the L2 a subsequent store hits on this shared block an...

Страница 170: ...Syntax Implementation Notes Return from Interrupt rfi rfi is context synchronizing which for the e500 means it works its way to the final execute stage updates architected registers and redirects the...

Страница 171: ...ation Notes TLB Invalidate Virtual Address Indexed tlbivax rA rB A TLB invalidate operation is performed whenever tlbivax is executed tlbivax invalidates any TLB entry that corresponds to the virtual...

Страница 172: ...if RA 0 then generate exception EA 32 0 GPR RB 32 63 ProcessID MAS6 SPID 0b0000_0000 AS MAS6 SAS VA AS ProcessID EA if Valid_TLB_matching_entry_exists VA then result see Table 12 15 column tlbsx hit e...

Страница 173: ...s are listed in Table 3 36 and Table 3 37 Table 3 31 Implementation Specific Instructions Summary Name Mnemonic Syntax Category TLB Invalidate Virtual Address Indexed tlbivax rA rB These are described...

Страница 174: ...uiring a cache access A word or half word memory access requires multiple accesses if it crosses a double word boundary but not if it crosses a natural boundary Vector loads and stores cause alignment...

Страница 175: ...y placing an isync instruction immediately following the loop containing the lwarx and stwcx The following example uses the Compare and Swap primitive to acquire the lock In this example it is assumed...

Страница 176: ...n instruction or sequence of instructions that prevents the store that releases a lock from being performed before stores caused by instructions preceding the barrier have been performed An export bar...

Страница 177: ...ds the store that releases the lock could be performed before those loads If it is necessary to ensure that those loads are performed before the store that releases the lock the programmer can either...

Страница 178: ...MO 1 the EIS defines mbar to function identically to the Classic PowerPC architecture definition of eieio If MO is not 1 the e500 executes mbar as though MO 0 The e500 core complex implements lwarx an...

Страница 179: ...registers as containing two 32 bit elements or four 16 bit elements as described in Section 3 8 1 3 SPE APU Instructions However like 32 bit Book E instructions scalar SPFP APU floating point instruct...

Страница 180: ...Machine state Single prec Single prec Interrupt Registers Single prec Single prec spr 62 ESR SPE Exception syndrome Single prec Single prec Interrupt Vector Offset Registers spr 405 IVOR5 Alignment A...

Страница 181: ...hese instructions can be found in the Instruction Set chapter of the EREF 3 8 1 1 SPE Operands Signed Fractions In signed fractional format the N bit operand is represented in a 1 N 1 format 1 sign bi...

Страница 182: ...identical for unsigned operands Table 3 34 shows how SPE APU vector multiply instruction mnemonics are structured Table 3 34 SPE APU Vector Multiply Instruction Mnemonic Structure Prefix Multiply Ele...

Страница 183: ...aaw Add to accumulator words Add word results to accumulator words pair of 32 bit sums an Add negated Add negated result to accumulator 64 bit sum anw Add negated to accumulator words Add negated word...

Страница 184: ...Vector Compare Less Than Unsigned evcmpltu crD rA rB Vector Convert Floating Point to Unsigned Integer with Round toward Zero evfsctuiz rD rB Vector Count Leading Sign Bits Word evcntlsw rD rA Vector...

Страница 185: ...dulo Integer evmhesmi rD rA rB Vector Multiply Half Words Even Signed Modulo Integer and Accumulate into Words evmhesmiaaw rD rA rB Vector Multiply Half Words Even Signed Modulo Integer and Accumulate...

Страница 186: ...e into Words evmhoumiaaw rD rA rB Vector Multiply Half Words Odd Unsigned Modulo Integer and Accumulate Negative into Words evmhoumianw rD rA rB Vector Multiply Half Words Odd Unsigned Modulo Integer...

Страница 187: ...tor Multiply Word Signed Saturate Fractional and Accumulate 3 evmwssfaa rD rA rB Vector Multiply Word Signed Saturate Fractional and Accumulate Negative 3 evmwssfan rD rA rB Vector Multiply Word Unsig...

Страница 188: ...Store Word of Word from Odd evstwwo rS d rA Vector Store Word of Word from Odd Indexed evstwwox rS rA rB Vector Subtract from Word evsubfw rD rA rB Vector Subtract Immediate from Word evsubifw rD UIM...

Страница 189: ...rt Floating Point from Unsigned Fraction efscfuf efdcfuf evfscfuf rD rB Convert Floating Point from Unsigned Integer efscfui efdcfui evfscfui rD rB Convert Floating Point Single from Double Precision...

Страница 190: ...int Test Less Than efststlt efdtstlt evfststlt crD rA rB SPE Double Word Load Store Instructions Vector Load Double Word into Double Word evldd rD d rA Vector Load Double Word into Double Word Indexed...

Страница 191: ...d the block is already in the L2 cache dcbtls marks the block so it is not a candidate for replacement Data Cache Block Touch for Store and Lock Set dcbtstls CT rA rB If CT 0 the e500 core fetches the...

Страница 192: ...n index in which all cache ways are already locked the least recently used way is evicted and L1CSR0 CLO is set indicating an overlock the new line is not locked or cached To precisely detect an overl...

Страница 193: ...the target instruction of the 512 most recently taken branches Table 3 43 lists the BTB instructions The branch buffer entry address register BBEAR and the branch buffer target address register BBTAR...

Страница 194: ...ess or if the entry exists but it is not locked the instruction is a no op and no other status is reported After bbelr executes the entry continues to be valid in the BTB with all its attributes uncha...

Страница 195: ...he address in the BBEAR the target address of that entry is overwritten with the target address in the BBTAR and BUCSR BBLO is set If all the ways of the BTB are locked for the index to which the BBEA...

Страница 196: ...k E PowerPC AIM e500 addc o evmwsmiaa SPE APU adde o evmwsmian SPE APU addi evmwssf SPE APU addic evmwssfa SPE APU addis evmwssfaa SPE APU addme o evmwssfan SPE APU add o evmwumi SPE APU addze o evmwu...

Страница 197: ...PU dcbtstls Cache locking extsb dcbz extsh divw o extsw 64 bit only divwu o fabs eciwx fadds ecowx fadd efdabs DPFP e500v2 fcfid efdadd DPFP e500v2 fcmpo efdcfs DPFP e500v2 efdcfsf DPFP e500v2 fcmpu e...

Страница 198: ...king efscmplt Scalar SPFP icbt efsctsf Scalar SPFP icbtls Cache locking efsctsi Scalar SPFP isel Integer select efsctsiz Scalar SPFP isync efsctuf Scalar SPFP lbz efsctui Scalar SPFP lbzu efsctuiz Sca...

Страница 199: ...PFP lwzux evfscfsi Vector SPFP lwzx evfscfuf Vector SPFP mbar evfscfui Vector SPFP mcrf evfscmpeq Vector SPFP mcrfs evfscmpgt Vector SPFP mcrxr evfscmplt Vector SPFP mfapidi evfsctsf Vector SPFP mfcr...

Страница 200: ...vlwhosx SPE APU mulldo evlwhou SPE APU mulli evlwhoux SPE APU mullw o evlwhsplat SPE APU nand evlwhsplatx SPE APU neg o evlwwsplat SPE APU nor evlwwsplatx SPE APU orc evmergehi SPE APU ori evmergehilo...

Страница 201: ...d evmhogsmiaa SPE APU stdcx evmhogsmian SPE APU stdu evmhogumiaa SPE APU stdux evmhogumian SPE APU stdx evmhosmf SPE APU stfd evmhosmfa SPE APU stfdu evmhosmfaaw SPE APU stfdux evmhosmfanw SPE APU stf...

Страница 202: ...bfze o evmwlsmiaaw SPE APU sync Replaced with msync Replaced with msync evmwlsmianw SPE APU tlbia evmwlssiaaw SPE APU tlbie evmwlssianw SPE APU tlbivax evmwlumi SPE APU tlbre evmwlumia SPE APU tlbsx e...

Страница 203: ...NOTE Some of these definitions differ slightly from those used to describe previous processors that implement the PowerPC architecture in particular with respect to dispatch issue finishing retiremen...

Страница 204: ...updated the architected registers Issue The stage responsible for reading source operands from rename registers and register files This stage also assigns instructions to the proper execution unit Lat...

Страница 205: ...ted in more than one stage simultaneously especially in the sense that a stage can be seen as a physical resource For example when instructions are dispatched they are assigned a place in the CQ at th...

Страница 206: ...d MU execution units are also multiple stage pipelines Figure 4 1 Instruction Flow Pipeline Diagram Showing Pipeline Stages Decode Stage SU1 Maximum four instruction BU BU SU2 fetch per clock cycle Fe...

Страница 207: ...ughput Instruction Cache Branch Issue Queue Decode Dispatch 2 instructions per clock cycle IQ9 IQ2 IQ1 IQ0 IQ11 IQ10 LSU 1 LSU 2 LSU 3 SU2 Signals completion queue when execution finishes CQ13 CQ1 CQ2...

Страница 208: ...FB shown in Figure 4 8 Likewise on a cache miss as many as four instructions can be forwarded to the fetch unit from the line fill buffer as the cache line is passed to the instruction cache Fetch tim...

Страница 209: ...rised of individual non blocking execution units implemented in parallel Each execution unit has a reservation station that must be available for an instruction issue to occur In most cases instructio...

Страница 210: ...sed for SPE and floating point vector instructions Only the lower half is used by scalar instructions including embedded single precision floating point instructions The execution unit executes the in...

Страница 211: ...tations The black stripe is a reminder that the instruction occupies an entry in the CQ described in Figure 4 4 Execute The operations specified by an instruction are being performed by the appropriat...

Страница 212: ...the IQ during each clock cycle Two instructions per clock cycle can be dispatched to the issue queues Two instructions from the GIQ and one instruction from the BIQ can issue per clock cycle to the a...

Страница 213: ...update either the LR or CTR write back their results in a similar fashion Section 4 3 1 General Instruction Flow describes this process 4 3 1 General Instruction Flow To resolve branch instructions an...

Страница 214: ...1 TLB arrays miss the access proceeds to the L2 TLB arrays For L1 instruction address translation misses the L2 TLB latency is at least 5 clocks for L1 data address translation misses the L2 TLB laten...

Страница 215: ...r the request for as many as four instructions to enter the IQ The cache is not blocked to internal accesses during a cache reload hits under misses The cache allows a hit under one miss and is only b...

Страница 216: ...gins during the same clock cycle that the rename register is updated with the data the instruction is dependent on The CQ maintains program order after instructions are dispatched guaranteeing in orde...

Страница 217: ...nd completion to one per cycle There are six basic types of instruction serialization Presync serialization Presync serialized instructions are held in the instruction queue until all prior instructio...

Страница 218: ...ction In general all stores and cache operation instructions are store serialized 4 3 4 Interrupt Latency The e500v1 flushes all instructions in the completion queue when an interrupt is taken except...

Страница 219: ...sactions caused by prior instructions complete entirely in its caches and externally on the bus address and data complete on the bus excluding instruction fetches No subsequent instructions and associ...

Страница 220: ...echanisms However unlike msync and mbar with MO 0 subsequent instructions can complete without waiting for mbar to perform its address bus tenure This provides a faster way to order data accesses 4 4...

Страница 221: ...resolve and complete until the cmp results are available add1 and add2 are dispatched to the GIQ In cycle 4 the bc resolves as correctly predicted add1 and add2 are issued to the SUs and are marked as...

Страница 222: ...ermines the fetching path based on the prediction and the target address If the prediction is wrong subsequent instructions and their results are purged Instructions ahead of the predicted branch proc...

Страница 223: ...oundary As shown in Figure 4 8 if instructions in a cache line are numbered 0 7 and the fetch group address maps to the nth instruction where n 0 1 2 3 or 4 instructions n n 1 n 2 n 3 are in the fetch...

Страница 224: ...ction in the fetch group The presence of instructions such as isync before the branch instruction Figure 4 9 shows all possible fetch group addresses FGAs that can be associated with a branch instruct...

Страница 225: ...gly taken is allocated for b2 at A for the fetch group address of A Later b1 is taken and thus mispredicted The BTB entry for address A becomes allocated for b1 replacing the prediction for b2 for the...

Страница 226: ...BUCSR This following BTB operations are controlled through BUCSR BTB disabling BUCSR BPEN is used to enable or disable the BTB The BTB is enabled when the bit is set and disabled when it is cleared W...

Страница 227: ...plies data to the GPRs by means of the LSU The core complex LSU is directly coupled to the data cache with a 64 bit 8 byte interface to allow efficient movement of data to and from the GPRs The LSU pr...

Страница 228: ...em LSU L1 load miss queue LMQ As loads reach the LSU it tries to access the cache On a hit the cache returns the data If there is a miss the LSU allocates an LMQ entry and a DLFB entry The LSU then qu...

Страница 229: ...Divide latency depends upon the operand data and ranges from 4 to 35 cycles as shown in Table 4 2 LSU data line fill buffer DLFB DLFB entries are used for loads and cacheable stores Stores are alloca...

Страница 230: ...cases 35 evdivwx Both the lower and upper words match the criteria described above for the divwx 4 cycle case 4 Assuming the 4 cycle evdivwx case does not apply the lower and upper words match the cr...

Страница 231: ...tage in clock cycle 10 Had divw2 been issued earlier it would have stalled in the reservation station until divw 1 vacated the second stage of the bypass path In other words the MU can hold as many as...

Страница 232: ...alignment of operands in memory may affect performance of memory accesses in some cases significantly as shown in Table 4 4 Alignment of memory operands on natural boundaries guarantees the best perfo...

Страница 233: ...4 3 describes branch instruction latencies Table 4 4 lists system operation instruction latencies The instructions in Table 4 4 are grouped by the serialization they require Except where otherwise no...

Страница 234: ...structions if marked with a 0 cycle execution time do not have an execute stage and all refetch serialized instructions have 1 cycle between the time they are completed and the time the target sequent...

Страница 235: ...nd BU 1 crandc BU 1 creqv BU 1 crnand BU 1 crnor BU 1 cror BU 1 crorc BU 1 crxor BU 1 mcrf BU 1 mcrxr BU 1 Presync postsync mfcr SU1 1 Move from mtcrf single field SU1 1 mtcrf multiple fields BU 2 Mov...

Страница 236: ...SU1 or SU2 1 1 extsh SU1 or SU2 1 1 isel SU1 or SU2 1 mulhwu MU 4 1 1 3 mulhw MU 4 1 1 3 mulli MU 4 1 3 mullw o MU 4 1 1 3 nand SU1 or SU2 1 1 neg o SU1 or SU2 1 1 nor SU1 or SU2 1 1 orc SU1 or SU2 1...

Страница 237: ...equent instruction can execute while CR results are generated 2 The MU provides a bypass path that allows divide instructions to perform the iterative operations necessary for division without blockin...

Страница 238: ...e evstddx 3 1 Store evstdh 3 1 Store evstdhx 3 1 Store evstdw 3 1 Store evstdwx 3 1 Store evstwhe 3 1 Store evstwhex 3 1 Store evstwho 3 1 Store evstwhox 3 1 Store evstwwe 3 1 Store evstwwex 3 1 Store...

Страница 239: ...resync lwbrx 3 1 lwz 3 1 lwzu 3 1 3 lwzux 3 1 3 lwzx 3 1 mbar 3 1 Store serialized msync Latency depends on bus response time Store and postsync serialized stb 3 1 Store stbu 3 1 3 Store stbux 3 1 3 S...

Страница 240: ...cycles may be needed for the data to reach the cache If the cache remains busy subsequent cache operations cannot execute 2 Section 4 3 3 3 Instruction Serialization describes the different types of...

Страница 241: ...tsf MU 4 1 efsctsi MU 4 1 efsctsiz MU 4 1 efsctuf MU 4 1 efsctui MU 4 1 efsctuiz MU 4 1 efsdiv MU 1 4 if either rA or rB is 0 0 29 all other cases efsmul MU 4 1 efsnabs SU1 or SU2 4 1 efsneg SU1 or SU...

Страница 242: ...ther the 4 or 11 cycle evdivwx cases apply the lower and upper words match the criteria described for the divwx 4 11 or 19 cycle case 1 All other cases1 1 eveqv SU1 evextsb SU1 1 evextsh SU1 1 evfsabs...

Страница 243: ...a MU 4 1 evmhesmfaaw MU 4 1 evmhesmfanw MU 4 1 evmhesmi MU 4 1 evmhesmia MU 4 1 evmhesmiaaw MU 4 1 evmhesmianw MU 4 1 evmhessf MU 4 1 evmhessfa MU 4 1 evmhessfaaw MU 4 1 evmhessfanw MU 4 1 evmhessiaaw...

Страница 244: ...4 1 evmhossianw MU 4 1 evmhoumi MU 4 1 evmhoumia MU 4 1 evmhoumiaaw MU 4 1 evmhoumianw MU 4 1 evmhousiaaw MU 4 1 evmhousianw MU 4 1 evmra MU 4 1 evmwhsmf MU 4 1 evmwhsmfa MU 4 1 evmwhsmi MU 4 1 evmwhs...

Страница 245: ...mwssfaa MU 4 1 evmwssfan MU 4 1 evmwumi MU 4 1 evmwumia MU 4 1 evmwumiaa MU 4 1 evmwumian MU 4 1 evnand SU1 1 evneg SU1 1 evnor SU1 1 evor SU1 1 evorc SU1 1 evrlw SU1 1 evrlwi SU1 1 evrndw SU1 1 evsel...

Страница 246: ...ched at a time Schedule instructions to minimize stalls due to busy execution units Avoid scheduling high latency instructions close together Interspersing single cycle latency instructions between lo...

Страница 247: ...when not all addresses are known at compile time or link time Because performance is typically not good position independent code should be avoided when possible 4 7 2 Dispatch Unit Resource Requirem...

Страница 248: ...7 3 Issue Queue Resource Requirements Instructions cannot be issued unless the specified execution unit is available The following sections describe how to optimize use of the issue queues 4 7 3 1 Gen...

Страница 249: ...Q0 is a mispredicted branch the entry in CQ1 should not be considered valid These and other less common conditions are described in the e500 Software Optimization Guide 4 7 4 1 Completion Groupings Th...

Страница 250: ...ion at the same time as an executing divide instruction As shown in Figure 4 1 and Figure 4 1 the MU consists of a multiply subunit and a divide subunit These subunits share the same reservation stati...

Страница 251: ...es cause alignment exceptions if they cross their natural alignment boundaries as show in Figure 4 9 Frequent unaligned accesses are discouraged because of the impact on performance Note the following...

Страница 252: ...Pipeline As shown in Figure 4 10 the e500v1 supports as many as four outstanding load misses in the load miss queue LMQ the e500v2 LMQ supports as many as nine Table 4 10 shows a load followed by a de...

Страница 253: ...ceptions as being generated by signals from internal and external peripherals instructions the internal timer facility debug events or error conditions There are three categories of interrupts describ...

Страница 254: ...ordered interrupt is taken See Section 5 10 Interrupt Ordering and Masking All interrupts except the machine check interrupt are context synchronizing as defined in the instruction model chapter of th...

Страница 255: ...is clear and software attempts to execute any of the SPE instructions the SPE unavailable interrupt is taken If this bit is set software can execute any SPE instructions NOTE On the e500v1 all SPFP in...

Страница 256: ...ion does not support IAC3 IAC4 DAC3 and DAC4 comparisons The core complex supports instruction address compare IAC1 and IAC2 and data address compare DAC1 and DAC2 for effective addresses only Real ad...

Страница 257: ...miss or data storage interrupt Interrupt vector prefix register IVPR IVPR 32 47 provides the high order 48 bits of the address of the interrupt handling routine for each interrupt type The 16 bit vec...

Страница 258: ...terrupt is taken the MCSR is updated to differentiate between machine check conditions Table 5 4 lists e500 bit assignments The MCSR also indicates whether a machine check condition is recoverable ABI...

Страница 259: ...ordering exception Data storage instruction storage 47 55 Reserved should be cleared 56 SPE SPE exception bit Book E allocates this bit for implementation dependent use so it may have different functi...

Страница 260: ...e program interrupt is taken Execution of a defined instruction using an invalid form illegal instruction exception type program interrupt unimplemented operation exception type program interrupt or p...

Страница 261: ...rupt handling routine is the address of the instruction that would have executed next had the asynchronous interrupt not occurred Synchronous Precise Caused directly by instruction execution Synchrono...

Страница 262: ...s Imprecise Imprecise interrupts may indicate the address of the instruction causing the exception that generated the interrupt or some instruction after that instruction When execution or attempted e...

Страница 263: ...VOR contents are indeterminate upon reset and must be initialized by system software Interrupts do not clear reservations obtained with load and reserve instructions The operating system should do so...

Страница 264: ...1 SP ST 6 Cache locking SP DLK ILK ST 7 Byte ordering SP ST BO IVOR3 Instruction storage ISI Access SP 5 20 Byte ordering SP BO IVOR4 External input A MSR EE 3 5 21 IVOR5 Alignment SP ST SPE ST 5 22 I...

Страница 265: ...or implementation specific ESR bits Legend xxx no brackets means ESR xxx is set xxx means ESR xxx could be set xxx yyy means either ESR xxx or ESR yyy may be set but never both xxx yyy means either ES...

Страница 266: ...MCSRR0 and MCSRR1 to save the return address and the MSR in case the machine check is recoverable Return From Machine Check Interrupt instruction rfmci is implemented to support the return to the add...

Страница 267: ...state immediately on detecting the machine check condition Table 5 8 e500 Machine Check Exception Sources Source Signal Additional Enable Bits Negative edge on machine check signal mcp mcp HID0 EMCP...

Страница 268: ...ache errors It describes error signaling and detection and it contains information about error recoverability The L1 caches in the e500 core complex are protected by parity Parity information is writt...

Страница 269: ...set For each condition the table provides comments about recoverability whether the MCAR has the address of the bad data whether the exception is precise and how far corrupted data can go into the GPR...

Страница 270: ...hen the field L1CSR1 ICPI will not be set Castout or snoop push BUS_WAERR Address bus error MCSRR0 is not meaningful MCAR is set to an address on the cache line with the error The system has enough in...

Страница 271: ...mplementation cannot access data in the byte order specified by the page s endian attribute Note The byte ordering exception is provided to assist implementations that cannot support dynamically switc...

Страница 272: ...ss of the instruction causing the interrupt SRR1 Set to the MSR contents at the time of the interrupt ESR ST Set if the instruction causing the interrupt is a store or store class cache management ins...

Страница 273: ...exists an external input exception is presented to the interrupt mechanism and MSR EE 1 The specific definition of an external input exception is implementation dependent and is typically caused by a...

Страница 274: ...defines the following alignment exception conditions Execution of a dcbz references a page marked as write through or cache inhibited A load multiple word instruction lmw reads an address that is not...

Страница 275: ...n an alignment interrupt occurs the processor suppresses the execution of the instruction causing the alignment exception SRR0 SRR1 MSR DEAR and ESR are updated as shown in Table 5 16 Instruction exec...

Страница 276: ...nvalid form boundedly undefined results On the e500 all instructions have invalid forms cause boundedly undefined results A reserved no op instruction no operation performed is preferred There are no...

Страница 277: ...ress IVPR 32 47 IVOR10 48 59 0b0000 MSR CE ME and DE are unchanged All other MSR bits are cleared ESR PIL Set if an illegal instruction exception type program interrupt otherwise cleared PPR Set if a...

Страница 278: ...h TCR FPEXT specifies one of 64 bit locations of the time base used to signal a fixed interval timer exception on a transition from 0 to 1 TCR FPEXT TCR FP 000000 selects TBU 32 TCR FPEXT TCR FP 11111...

Страница 279: ...Writing a 1 to this bit causes it to be cleared writing a 0 has no effect 5 7 12 Data TLB Error Interrupt A data TLB error interrupt occurs when no higher priority exception exists and the exception d...

Страница 280: ...nts at the time of the interrupt MSR CE ME and DE are unchanged All other MSR bits are cleared DEAR Set to the EA of a byte that is both within the range of the bytes being accessed by the memory acce...

Страница 281: ...nd MMU Background chapter of the EREF describes how these values are set as defined by the EIS Instruction execution resumes at address IVPR 32 47 IVOR14 48 59 0b0000 WD ID MD GD ED SPID0 PID0 SAS MSR...

Страница 282: ...errupts are enabled DBCR0 IDM 1 and MSR DE 1 CSRR0 is set as follows For instruction address compare IAC registers data address compare DAC1R DAC1W DAC2R and DAC2W trap TRAP or branch taken BRT debug...

Страница 283: ...cleared and an SPE embedded scalar double precision e500v2 only or embedded vector single precision floating point instruction is executed It is not used by the embedded scalar single precision float...

Страница 284: ...ettings Instruction execution resumes at address IVPR 32 47 IVOR33 48 59 0b0000 5 7 15 3 Embedded Floating Point Round Interrupt The embedded floating point round interrupt is taken on any of the foll...

Страница 285: ...errupt condition could occur with MSR EE 0 the interrupt cannot be taken until MSR EE 1 If a counter overflows while PMGC0 FCECE 0 PMLCan CE 1 and MSR EE 0 it is possible for the counter to wrap aroun...

Страница 286: ...ny load bytes at the addressed location may have beenaccessed if read access to thatpage in which bytes were accessed is permitted by the access control mechanism For load multiple or load string some...

Страница 287: ...hich is edge triggered All asynchronous interrupt types and some synchronous interrupt types can be masked The PowerPC architecture allows implementations to avoid situations in which an interrupt wou...

Страница 288: ...synchronous interrupts as well as in the case of MSR DE any debug interrupts including synchronous and asynchronous types Branching or sequential execution to addresses not mapped by the TLB mapped wi...

Страница 289: ...5 noncritical interrupts are immediately followed by the highest priority existing critical interrupt type without executing any instructions at the noncritical interrupt handler This is because noncr...

Страница 290: ...ermitted setting of multiple exceptions for which the corresponding interrupt types are disabled The generation of exceptions for which the corresponding interrupt types are disabled has no effect on...

Страница 291: ...truction address compare 8 ITLB miss 9 ISI 10 SPE embedded floating point APU unavailable 11 Program 12 DTLB miss 13 DSI 14 Alignment 15 Embedded floating point data interrupt 16 Embedded floating poi...

Страница 292: ...d store tag when lac_clear_guarded is asserted the guarded attribute in CQ0 is cleared This process allows the completion unit to take an interrupt If a cache inhibited stwcx gets an address error the...

Страница 293: ...e logic to initiate actions that cause the core complex to enter core halted state as follows Suspend instruction fetching Complete all previously fetched instructions When the instruction pipeline is...

Страница 294: ...logic may similarly stop clocking to idle device level blocks tben I Asserted by the integrated device logic to enable the time base tbint O Asserted when a time base interrupt is signaled This ordin...

Страница 295: ...input which suspends timer facility operations Note that tben controls the time base and decrementer in all power management states Timer operation is independent of power management except for softwa...

Страница 296: ...power saving state doze nap or sleep has been previously set up by setting the appropriate HID0 bit typically at system start up time Setting MSR WE does not directly affect instruction execution but...

Страница 297: ...ould typically respond to doze nap and sleep and control the core s power state through the halt halted stop stopped and tben tbint signal pairs Figure 6 2 Example Core Power Management Handshaking MS...

Страница 298: ...he handler issues an rfi rfci or rfmci As a result doze nap and sleep outputs negate to the external power management logic on entry to the interrupt service routine and then return to their previous...

Страница 299: ...s mispredicted branches or the number of cycles an execution unit stalls The count of such events can be used to trigger the performance monitor interrupt The performance monitor can be used to do the...

Страница 300: ...ned to interrupt vector offset register 35 IVOR35 Its priority is less than the fixed interval interrupt and greater than the decrementer interrupt Software communication with the performance monitor...

Страница 301: ...tor global control 0 PMGC0 Table 7 2 Performance Monitor Registers User Level Read Only Number PMR 0 4 PMR 5 9 Name Abbreviation 0 00000 00000 Performance monitor counter 0 UPMC0 1 00000 00001 Perform...

Страница 302: ...t value until it is changed by software 0 The PMCs are incremented if permitted by other PM control bits 1 The PMCs are not incremented 33 PMIE Performance monitor interrupt enable 0 Performance monit...

Страница 303: ...ors in the system Time base frequency is implementation dependent so software should invoke a system service program to obtain the frequency before choosing a TBSEL value 53 54 Reserved should be clea...

Страница 304: ...ounter in supervisor state 0 The PMC can be incremented if enabled by other performance monitor control fields 1 The PMC cannot be incremented if MSR PR is cleared 34 FCU Freeze counter in user state...

Страница 305: ...ld is multiplied by 1 PMLCbn THRESHOLD 1 001 Threshold field is multiplied by 2 PMLCbn THRESHOLD 2 010 Threshold field is multiplied by 4 PMLCbn THRESHOLD 4 011 Threshold field is multiplied by 8 PMLC...

Страница 306: ...interrupt is masked for many cycles during which the counters may continue incrementing A high order bit is set normally only when the counter increments from a value below 2 147 483 648 0x8000_0000 t...

Страница 307: ...s clear until after the counter leaves the overflow state msb becomes 0 or if EE remains clear until after PMLCan CE or PMGC0 PMIE cleared the exception is not signaled The following sequence is recom...

Страница 308: ...s clear and freezing of the counters is not enabled PMGC0 FCECE is clear PMCn can wrap around to all zeros again without the performance monitor interrupt being taken 7 5 Event Counting This section d...

Страница 309: ...nter control or by setting PMLCan FCS and PMLCan FCU for each counter control 7 6 Examples The following sections provide examples of how to use the performance monitor facility 7 6 1 Chaining Counter...

Страница 310: ...l buffer DLFB load miss cycles event C0 76 and C1 76 require a threshold value A DLFB load miss cycles event is counted only when the number of cycles spent recovering from the miss is greater than th...

Страница 311: ...2 for load store with update 1 32 for load or store multiple instructions Com 4 Instructions fetched Spec Fetched instructions 0 1 2 3 or 4 per cycle instructions written to the IQ Com 5 Micro ops de...

Страница 312: ...Cycles MU is not empty but 0 instructions scheduled Com 24 Cycles LRU schedule stalled Spec Cycles LRU is not empty but 0 instructions scheduled Com 25 Cycles BU schedule stalled Spec Cycles BU is not...

Страница 313: ...c Counts number of stalls Com 55 counts cycles stalled Com 48 Data MMU miss Spec Counts number of stalls Com 56 counts cycles stalled Com 49 Data MMU busy Spec Counts number of stalls Com 57 counts cy...

Страница 314: ...ide requests Spec Master instruction side assertions of ts Com 69 BIU master data side requests Spec Master data side assertions of ts Com 70 BIU master data side castout requests Spec Includes replac...

Страница 315: ...ng fetch to the critical interrupt vector exceeds the threshold See note for previous event Chaining Events3 Com 82 PMC0 overflow N A PMC0 32 transitions from 1 to 0 Com 83 PMC1 overflow N A PMC1 32 t...

Страница 316: ...PowerPC e500 Core Family Reference Manual Rev 1 7 18 Freescale Semiconductor Performance Monitor...

Страница 317: ...he Book E PowerPC architecture 8 2 Programming Model This section describes the registers instructions and interrupts defined by the Book E architecture to support the debug facility 8 2 1 Register Se...

Страница 318: ...these registers to restore state at the end of the interrupt handler Debug interrupts do not affect the save restore registers SRR0 and SRR1 and CSRR registers are not affected by the Return from Int...

Страница 319: ...1R DAC1W DAC2R and DAC2W debug events trap or branch taken debug exceptions set to the address of the instruction causing the debug interrupt For instruction complete debug exceptions set to the addre...

Страница 320: ...roup is a serial protocol that specifies data flow though special registers connected between test data in TDI and test data out TDO Figure 8 1 shows the TAP registers implemented by the core complex...

Страница 321: ...low the system designer to disable test functions that would impede normal operation Test reset TRST is an optional JTAG signal used in the e500 to reset the TAP controller asynchronously This signal...

Страница 322: ...State Meaning Asserted Negated The contents of the selected internal instruction or data register are shifted out onto this signal Valid data appears on the falling edge of TCK Quiescent except when...

Страница 323: ...Address Compare Debug Event One or more instruction address compare debug events IAC1 and IAC2 occur if they are enabled and execution is attempted of an instruction at an address that meets the crit...

Страница 324: ...for detecting IAC register debug events Instruction address compare debug events can occur regardless of the values of MSR DE or DBCR0 IDM When an instruction address compare debug event occurs the c...

Страница 325: ...debug events can occur on read type data accesses and whether DAC1W debug events can occur on write type data accesses DBCR0 DAC2 specifies whether DAC2R debug events can occur on read type data acces...

Страница 326: ...ies the following Whether all or some of the address bits for the data access must match the contents of DAC1 or DAC2 Whether the address must be inside or outside of a range specified by DAC1 and DAC...

Страница 327: ...n this case CSRR0 contains the address of the instruction following the instruction that enabled the debug interrupt The debug interrupt handler can observe DBSR IDE to determine how to interpret the...

Страница 328: ...ebug event occurs when any instruction completes execution so long as MSR DE and DBCR0 ICMP are both set instruction complete debug events are enabled Note that no instruction complete debug event occ...

Страница 329: ...tical interrupt vector that caused the event No instructions at the noncritical interrupt handler are executed If debug interrupts are disabled when the event occurs no interrupt is generated However...

Страница 330: ...lements the e500 core for details An unconditional debug event can occur regardless of the value of MSR DE and is the only debug event that does not have a corresponding debug control register enable...

Страница 331: ...Caches describes the organization of the on chip level one instruction and data caches cache coherency protocols cache control instructions and various cache operations It describes the interaction th...

Страница 332: ...PowerPC e500 Core Family Reference Manual Rev 1 Part II 2 Freescale Semiconductor...

Страница 333: ...r a specified amount of time has elapsed unless one of the following occurs DEC is altered by software in the interim The TB update frequency changes The DEC is typically used as a general purpose sof...

Страница 334: ...mer facilities HID0 Clock source select and enable The clock source for the core timer facilities is specified by two fields in the hardware implementation dependent register 0 HID0 time base enable T...

Страница 335: ...ase is static there is no counting If the time base is enabled HID0 TBEN is set the clock source is determined as follows If HID0 SEL_TBCLK 0 the timer facilities are updated every 8 CCB clocks If HID...

Страница 336: ...the e500v2 this frequency is the core frequency The ATBU and ATBL registers can be read by executing an mfspr instruction but cannot be written Reading the ATB or ATBL register places the lower 32 bit...

Страница 337: ...e500 supports the following APUs defined by the EIS Integer select APU Performance monitor APU Signal processing engine APU SPE APU Embedded floating point APUs Embedded vector single precision floati...

Страница 338: ...e BTB provides quick access to branch targets and history bits that allow efficient branch prediction The core complex also provides support for locking and unlocking BTB entries for deterministic bra...

Страница 339: ...PU The alternate time base APU defines a time base counter similar to the time base defined in PowerPC architecture It is intended to be used for measuring time in implementation defined intervals It...

Страница 340: ...uble precision floating point operations use the SPEFSCR as it is described in the EREF Double precision floating point instructions affect only the low element floating point status flags and leave t...

Страница 341: ...FFFF_FFFF which is approximately 1 8E 307 and the minimum positive normalized value pmin is represented by the encoding 0x0010_0000_0000_0000 approximately 2 2E 308 Biased exponent values 0 and 2047 a...

Страница 342: ...iplication if Aexp Bexp 1023 2046 then overflow An underflow is said to have occurred if the numerically correct result of an instruction is such that 0 r pmin or nmin r 0 In this case r may be denorm...

Страница 343: ...nmax as appropriate is stored in rD If underflow occurs 0 for rounding modes RN RZ RP or 0 for rounding mode RM is stored in rD Exceptions If the contents of rA or rB are Infinity Denorm or NaN SPEFS...

Страница 344: ...a32Denorm f then SPEFSCRFINV 1 result fsign 63 0 else resultsign fsign resultexp fexp 127 1023 resultfrac ffrac 29 0 rD0 63 result The single precision floating point value in the low element of rB is...

Страница 345: ...32 63 SIGN I The signed integer low element in rB is converted to a double precision floating point value using the current rounding mode and the result is placed into rD Exceptions None efdcfuf efdcf...

Страница 346: ...1 else cl 0 CR4 crD 4 crD 3 undefined cl undefined undefined rA is compared against rB If rA is equal to rB the bit in the crfD is set otherwise it is cleared Comparison ignores the sign of 0 0 0 Exce...

Страница 347: ...nfinities and Denorms as normalized numbers using their values of e and f directly efdcmplt efdcmplt Floating Point Double Precision Compare Less Than efdcmplt crfD rA rB al rA0 63 bl rB0 63 if al bl...

Страница 348: ...sult and FG and FX are updated so the handler can perform rounding efdctsi efdctsi Convert Floating Point Double Precision to Signed Integer efdctsi rD rB rD32 63 CnvtFP64ToI32Sat rB0 63 SIGN ROUND I...

Страница 349: ...D is updated with the truncated result and FG and FX are updated so the handler can perform rounding efdctuf efdctuf Convert Floating Point Double Precision to Unsigned Fraction efdctuf rD rB rD32 63...

Страница 350: ...FG and FX are updated so the handler can perform rounding efdctuiz efdctuiz Convert Floating Point Double Precision to Unsigned Integer with Round toward Zero efdctuiz rD rB rD32 63 CnvtFP64ToI32Sat r...

Страница 351: ...rB is 0 and the content of rA is a finite normalized non zero number SPEFSCR FDBZ is set If floating point divide by zero Exceptions are enabled an interrupt is then taken Otherwise if an overflow oc...

Страница 352: ...is set an interrupt is taken and rD is not updated Otherwise if an overflow occurs SPEFSCR FOVF is set or if an underflow occurs SPEFSCR FUNF is set If either underflow or overflow exceptions are enab...

Страница 353: ...for a normalized number If SPEFSCR FINVE 1 an interrupt is taken and rD is not updated efdneg efdneg Floating Point Double Precision Negate efdneg rD rA rD0 63 rA0 rA1 63 The sign bit of rA is complem...

Страница 354: ...et If SPEFSCR FINVE is set an interrupt is taken and rD is not updated Otherwise if an overflow occurs SPEFSCR FOVF is set or if an underflow occurs SPEFSCR FUNF is set If either underflow or overflow...

Страница 355: ...trict IEEE 754 compliance is required the program should use efdcmpeq efdtstgt efdtstgt Floating Point Double Precision Test Greater Than efdtstgt crfD rA rB al rA0 63 bl rB0 63 if al bl then cl 1 els...

Страница 356: ...efined rA is compared against rB If rA is less than rB the bit in the crfD is set otherwise it is cleared Comparison ignores the sign of 0 0 0 The comparison proceeds after treating NaNs Infinities an...

Страница 357: ...rB is converted to a single precision floating point value using the current rounding mode and the result is placed into the low element of rD Exceptions If the rB value is Infinity Denorm or NaN SPEF...

Страница 358: ...malized positive number The encoding for double precision is 0x0010_0000_0000_0000 nmin denotes the minimum normalized negative number The encoding for double precision is0x8010_0000_0000_0000 10 4 5...

Страница 359: ...p fpsign 0b11111110 231 else fpexp fpexp 1 fpfrac v1 23 else fpfrac v1 23 else if SPEFSCRFRMC 0b10 0b10 then infinity modes implementation dependent return fp Round a result Round64 fp guard sticky FP...

Страница 360: ...f fpexp 0 fpfrac 0 then return 0x00000000 all zero values if fractional I then convert to integer max_exp 1054 shift 1054 fpexp if signed SIGN then if fpexp 1054 fpfrac 0 fpsign 1 then max_exp max_exp...

Страница 361: ...uration Convert from integer fractional to 64 bit floating point signed SIGN or UNSIGN fractional F fractional or I integer CnvtI32ToFP64Sat v signed fractional FP64format result resultsign 0 if v 0 t...

Страница 362: ...PowerPC e500 Core Family Reference Manual Rev 1 10 26 Freescale Semiconductor Auxiliary Processing Units APUs...

Страница 363: ...hat can access system memory maintain their own caches and function as bus masters requiring cache coherency 11 1 Overview The core complex L1 cache implementation has the following characteristics Se...

Страница 364: ...ith respect to the data cache Both instruction and data cache lines are filled in a single cycle 32 byte write from line fill buffers as described in Section 11 1 1 1 Load Store Unit LSU and Section 1...

Страница 365: ...ignment to and from the data cache provides sequencing for load store multiple operations and interfaces with the core interface unit Write operations to the data cache can be performed on a byte half...

Страница 366: ...v2 and the three entry data line fill buffer DLFB five entry in the e500v2 see Section 4 4 2 1 Load Store Unit Queueing Structures The LSU then queues a bus transaction to read the line If a subsequen...

Страница 367: ...outs and one that can be used for either 11 1 1 2 Instruction Unit The instruction unit interfaces with the L1 instruction cache and the core interface unit When instructions miss in the instruction c...

Страница 368: ...rovide the index to select a cache set The tags consist of physical address bits PA 0 19 Address translation occurs in parallel with set selection from PA 20 26 Lower address bits PA 27 31 locate a by...

Страница 369: ...t and PA 27 28 selects an instruction within a block The tags consist of physical address bits PA 0 19 Address translation occurs in parallel with set selection from PA 20 26 The instruction cache can...

Страница 370: ...L1 caches whenever one of the following occurs A store instruction or dcbz or dcba modifies the data cache A line fill occurs into the instruction or data cache L1 cache parity is checked whenever A...

Страница 371: ...L1CSR1 ICPI is set Similarly for the data cache L1CSR0 CPE must be set if L1CSR0 CPI is set If the programmer attempts to set L1CSR0 CPI using mtspr without setting L1CSR0 CPE then L1CSR0 CPI will no...

Страница 372: ...respect to main memory It does not reside in any other coherent caches 100 Exclusive E The line is present in the cache and this cache has exclusive ownership of the line It is not present in any oth...

Страница 373: ...ts only invalid and valid state Table 11 3 describes how execution of instruction cache control instructions affect L1 instruction cache coherency states dcbtstls CT 1 00x Any I dcbz 00x Any M icblc C...

Страница 374: ...snooped by other bus masters To determine the action to take due to a snoop the cache coherency protocol uses transfer type ttx encodings which are transmitted on the CCB with the address See Section...

Страница 375: ...gh and the other mapped as write back that is non write through In this case the cache line remains in its current state the store data is written into the data cache and the store goes to the CCB as...

Страница 376: ...an be recognized when a guarded load is in progress so the above precautions do not apply 11 3 5 Load Store Operations Load and store operations are assumed to be weakly ordered on the core complex Th...

Страница 377: ...they do not store to overlapping bytes of data Note that although memory accesses that miss in the L1 cache are forwarded onto the core interface unit for future arbitration onto the CCB all potential...

Страница 378: ...on specific information The remainder of this section describes how the cache control instructions and the L1CSRn bits are used to control the L1 cache 11 4 1 Cache Control Instructions The following...

Страница 379: ...r L1CSR0 CUL Acronyms are used to signify the following interrupts DTLB data TLB interrupt ALI alignment interrupt DSI data storage interrupt icbt Instruction Cache Touch no op x icbtls Instruction Ca...

Страница 380: ...nored and the corresponding cache is not accessed The default power up state of L1CSR0 CE and L1CSR1 ICE is zero caches disabled When the data cache is disabled snooping of lines in the cache is not p...

Страница 381: ...e that invalidating the caches resets all cache status bits including lock bits Also note that with dcbi the e500 core invalidates the cache block without pushing it out to memory See Section 3 3 1 8...

Страница 382: ...ll for a new address in a completely locked cache set will not be put into the cache It is however loaded into a DWB and creates the appropriate normal burst write transfer The cache locking DSI handl...

Страница 383: ...owing cache instructions are treated as stores and may cause the invalidation and unlocking of a cache line in another processor in a multiprocessor system dcba dcbz In implementations with an L2 cach...

Страница 384: ...ion then execute dcbf instructions to that region Note that a 48 Kbyte region must be used to ensure that the PLRU algorithm flushes all of the cache entries 12 x 128 sets x 32 bits 48 Kbytes Perform...

Страница 385: ...not loaded into the data cache If the data is modified it is loaded into a DWB and creates the appropriate normal burst write transfer Each of the eight ways of each set in the instruction cache can b...

Страница 386: ...not cause a fill Also cache operations such as dcbi and dcbf that miss in the cache do not cause a fill 11 6 1 4 Store Miss Merging When a caching allowed store misses in the data cache an entry is a...

Страница 387: ...L1 data cache function similarly to L1 instruction cache misses They cause a new line to be allocated on a PLRU basis provided the cache is not completely locked or disabled Note that modified data i...

Страница 388: ...set are updated using the rules specified in Table 11 9 Note that only three PLRU bits are updated for any access Table 11 9 PLRU Bit Update Rules Current Access New State of the PLRU Bits B0 B1 B2 B3...

Страница 389: ...B the core complex provides an address lock attribute CL on the bus which can be used in conjunction with the transfer type ttx encodings to identify which addresses to lock or unlock When the core co...

Страница 390: ...2 cache may also use other bus transactions to cause locks to be cleared such as bus transactions as a result of dcbf identified on the bus as an address only FLUSH or as an L1 push due to dcbf 11 7 2...

Страница 391: ...l 2 L2 MMU The L1 MMUs are completely invisible with respect to the architecture The programming model for implementing translation lookaside buffers TLBs provided in Book E and the Freescale Book E s...

Страница 392: ...es resident in L2 MMU completely maintained by the hardware Automatically performs invalidations to maintain consistency with L2 TLBs The level 2 MMU has the following features A 16 entry fully associ...

Страница 393: ...y instruction 12 4 1 12 18 tlbwe TLB Write Entry instruction 12 4 2 12 19 tlbsx rA rB preferred form is tlbsx 0 rB TLB Search for entry instruction 12 4 3 12 19 tlbivax rA rB TLB Invalidate entries in...

Страница 394: ...cant 12 bits always index within the page and are untranslated The appropriate L1 MMU instruction or data is checked for a matching address translation first If it misses the request for translation i...

Страница 395: ...respectively The e500 constructs three virtual addresses for each access The core complex implements three process ID PID registers PID0 PID2 as SPRs shown in Section 2 12 1 Process ID Registers PID0...

Страница 396: ...ddress spaces 12 2 2 Variable Sized Pages There are two kinds of TLBs on the e500 core complex as follows TLBs that translate addresses for 4 Kbyte pages only These TLBs are set associative based on t...

Страница 397: ...ess A hit to multiple matching TLB entries is considered a programming error If this occurs the TLB generates an invalid address and TLB entries may be corrupted an exception is not reported Figure 12...

Страница 398: ...ative TLB0 in the e500v2 Figure 12 4 Two Level MMU Structure Additionally Figure 12 4 shows that when the L2 MMU is checked for a TLB entry both TLB1 and TLB0 are checked in parallel It also identifie...

Страница 399: ...accesses both the L1TLB4K and the L1VSP TLBs in the instruction MMU are searched in parallel for the matching TLB entry Similarly for data accesses both the L1TLB4K and the L1VSP TLBs in the data MMU...

Страница 400: ...U replacement algorithm The LRU bits are updated each time a TLB entry is accessed for translation However there are other speculative accesses performed to the L1 MMUs that cause the LRU bits to be u...

Страница 401: ...associative array that supports nine e500v1 or eleven e500v2 page sizes TLB0 a 256 entry 2 way e500v1 or 512 entry 4 way e500 v2 set associative array that supports only 4 Kbyte page sizes The two L2...

Страница 402: ...cated internal case by an external tlbivax instruction or by a flash invalidate initiated by writing to the MMUCSR0 The IPROT bit can be used to protect critical code and data such as interrupt vector...

Страница 403: ...implemented completely by the system software Thus when an entry in TLB1 is to be replaced the software selects which entry to replace and writes the entry number to the MAS0 ESEL field before execut...

Страница 404: ...re automatically loads the current value of TLB0 NV1 into MAS0 ESEL and the complement of TLB0 NV into MAS0 NVlsb This sets up MAS0 such that if those values are not overwritten the alternate way will...

Страница 405: ...reloaded using entries from their level 2 array equivalent For example if the L1 data MMU misses but there is a hit for one of the three virtual addresses in TLB1 the matching entry is automatically l...

Страница 406: ...come inaccessible due to the snooping activity caused by the TLBINV 12 3 5 The G Bit of WIMGE The G bit provides protection from bus accesses due to speculative and faultable instruction execution A s...

Страница 407: ...ad Entry tlbre instruction and data is written to the TLBs from the MAS registers with a TLB Write Entry tlbwe instruction Table 12 4 TLB Entry Bit Definitions for e500 Field Comments V Valid bit for...

Страница 408: ...MAS3 and MAS7 if HID0 EN_MAS7_UPDATE 1 result Note that architecturally if the instruction specifies a TLB entry that is not found the results placed in MAS0 MAS3 and optionally MAS7 are undefined Ho...

Страница 409: ...and then executing the tlbwe instruction To write an entry into TLB1 MAS0 TLBSEL must 01 and MAS0 ESEL must point to the desired entry When the tlbwe instruction is executed the TLB entry information...

Страница 410: ...is especially useful for finding the TLB entry that caused a DSI or ISI exception In this case at most three tlbsx instructions are required one for each of the current PID values Note that TID value...

Страница 411: ...he core complex always snoops TLB invalidate transactions from other CCB bus masters if any and invalidates matching TLB entries accordingly Note that entries in TLB1 can be protected from invalidatio...

Страница 412: ...sync instruction has pending memory accesses that were issued before any previous tlbivax instructions were completed This instruction effectively synchronizes the invalidation of TLB entries tlbsync...

Страница 413: ...ore information on some of the actions taken by Freescale Book E devices on MMU exceptions The following subsections provide supplementary information that applies for the e500 12 5 1 Automatic Update...

Страница 414: ...E devices and for the e500 is that the e500 only uses MAS6 SPID0 and the e500 does not implement MAS5 Note that for a permissions violation case software must explicitly load a value into MAS6 SPID0 t...

Страница 415: ...ces to other sections that have more detailed bit descriptions for the e500 registers related to the MMU Also the EREF lists the Freescale Book E definitions for these registers Table 12 6 TLB1 Entry...

Страница 416: ...ts 46 47 that identifies which of the indexed entries is to be referenced by the TLB operation ESEL selects the way For TLB1 ESEL selects one of the 16 entries in the array Figure 12 11 describes the...

Страница 417: ...s field is updated based on the calculated next victim value for TLB0 based on the round robin replacement algorithm described in Section 12 3 2 2 Replacement Algorithms for L2 MMU Note that for the e...

Страница 418: ...Mbyte 1010 1 Gbyte 1011 4 Gbyte 56 63 Reserved should be cleared SPR 626 Access Supervisor only 32 51 52 56 57 58 59 60 61 62 63 R EPN X0 X1 W I M G E W Reset All zeros Figure 12 13 MAS Register 2 MA...

Страница 419: ...ian which differs from the modified little endian byte ordering model optionally available in previous devices that implement the original PowerPC architecture 0 The page is accessed in big endian byt...

Страница 420: ...e current PID registers should be used to load the MAS1 TID field on a TLB miss exception The e500 implementation defines this field as follows 00 PID0 01 PID1 10 PID2 11 TIDZ 0x00 all zeros 48 51 Res...

Страница 421: ...0 EN_MAS7_UPDATE Figure 12 17 shows the format of the MAS7 register The MAS7 fields are described in Table 12 14 SPR 630 Access Supervisor only 32 39 40 47 48 62 63 R SPID0 SAS W Reset All zeros Figur...

Страница 422: ...TLBSELD 0 TLB0 NV else undefined if TLBSEL 0 TLB0 NV else undefined if TLBSELD 0 TLB0 NV else undefined if TLBSEL 0 TLB0 NV else undefined V 1 1 0 V array IPROT 0 Matched IPROT if TLB1 hit else 0 0 I...

Страница 423: ...defined as follows High speed on chip local bus interface 32 bit address bus Address protocol with address pipelining and retry copyback derived from bus used by previous generations of PowerPC proces...

Страница 424: ...inhibit Normally reflected from the I bit of the WIMGE bits regardless of whether the cache is enabled For burst writes and address only transactions ci is always negated cl O Cache lock Indicates L2...

Страница 425: ...500 clock Refer to Chapter 8 Debug Support ckstp_out O Checkstop interrupt Assertion of this signal by the e500 core is used by system to generate a chip wide hard stop and to signal an external CKSTP...

Страница 426: ...then halt so the interrupt can be serviced wrs 0 1 O Watchdog timer reset status These two bits are set to one of three values when a reset is caused by the watchdog timer These bits are undefined at...

Страница 427: ...after the core complex has entered halted state otherwise the negation may not be recognized stop I Asserted by system logic to request that the core complex go from the halted state into the power d...

Страница 428: ...ctions and associated memory transactions are initiated until such completion occurs Execution of msync also generates a SYNC command on the bus if HID1 ABE is set through the tt 0 4 signals which als...

Страница 429: ...lex in certain clock modes to process a snoop transaction Note that address streaming as defined here differs from address pipelining which is the issue of multiple address tenures independent of whet...

Страница 430: ...ck the specified address from its cache This transaction is always performed as non global because it is specifically targeted at an L2 cache An L2 cache may also use other bus transactions to cause l...

Страница 431: ...cx behavior defined by the PowerPC architecture such use would have to be carefully controlled by the system 13 7 Remote Atomic Status Monitoring For system convenience the core complex provides a sys...

Страница 432: ...use of the machine check This is the only instance where RFXE should be set except for the case for the e500v1 described in the HID1 RFXE bit description of Section 2 10 2 Hardware Implementation Depe...

Страница 433: ...how the retry strategy can affect performance Load and reserve and store conditional instructions depend on the coherence mechanism of the system Stores to a given location are coherent if they are se...

Страница 434: ...show how the lwarx and stwcx instructions can be used to implement various synchronization primitives The sequences used to emulate the various primitives consist primarily of a loop using lwarx and s...

Страница 435: ...s of the word to be incremented is in GPR3 the increment is in GPR4 and the old value is returned in GPR5 loop lwarx r5 0 r3 load and reserve add r0 r4 r5 increment word stwcx r0 0 r3 store new value...

Страница 436: ...arison In this example it is assumed that the address of the word to be tested is in GPR3 the comparand is in GPR4 and the old value is returned there and the new value is in GPR5 loop lwarx r6 0 r3 l...

Страница 437: ...e of Section A 1 1 Synchronization Primitives thereby atomically loading the old value of the lock writing to the lock the new value 1 given in GPR4 returning the old value in GPR5 not used below and...

Страница 438: ...ement pointer is at offset 0 from the start of the element It is also assumed that the next element pointer of each list element is in a reservation granule separate from that of the next element poin...

Страница 439: ...ocessor is implementation dependent see Section 3 3 1 7 Atomic Update Primitives Using lwarx and stwcx In some implementations performance may be improved by minimizing looping on a lwarx instruction...

Страница 440: ...PowerPC e500 Core Family Reference Manual Rev 1 A 8 Freescale Semiconductor Programming Examples...

Страница 441: ...64 Bit Specific Book E Instructions A subset of Book E instructions are restricted to 64 bit Book E processing A 32 bit Book E implementation need not implement any of the following instructions Like...

Страница 442: ...eed to be calculated and presented to main memory Given that only branch and data memory access instructions not included in Section B 1 64 Bit Specific Book E Instructions are defined to prepend 32 z...

Страница 443: ...produce the full 64 bit address Book E also provides a complete set of branch instructions that perform a modulo 232 on the computed branch target effective address and then prepend 32 zeros to produc...

Страница 444: ...PowerPC e500 Core Family Reference Manual Rev 1 B 4 Freescale Semiconductor Guidelines for 32 Bit Book E...

Страница 445: ...mer to program using more intuitive mnemonics and symbols than the instructions and syntax defined by the instruction set architecture For example to code the conditional call branch to an absolute ta...

Страница 446: ...be difficult to understand Simplified mnemonics are provided for the following operations Extract Select a field of n bits starting at bit position b in the source register left or right justify this...

Страница 447: ...left 8 bits slwi rA rA 8 equivalent to rlwinm rA rA 8 0 23 4 Clear the high order 16 bits of rS and place the result into rA clrlwi rA rS 16 equivalent to rlwinm rA rS 0 16 31 Table C 3 Word Rotate an...

Страница 448: ...s for BO and BI For example bc 16 0 target is a conditional branch that as a BO value of 16 0b1_0000 indicates decrements the CTR then branches if the decremented CTR is not zero The operation specifi...

Страница 449: ...Mnemonics The following key points are helpful in understanding how to use simplified branch mnemonics All simplified branch mnemonics eliminate the BO operand so if any operand is present in a branc...

Страница 450: ...of the Instruction Encoding Table C 5 BO Bit Encodings BO Bit Description 0 If set ignore the CR bit comparison 1 If set the CR bit comparison is against true if not set the CR bit comparison is agai...

Страница 451: ...ds coding the suffix causes the bit to be set and coding the suffix causes the bit to be cleared For branches to an address in the LR or CTR bclr l or bcctr l coding the suffix causes the y bit to be...

Страница 452: ...used to replace the decimal operand as shown in the example in Section C 4 Branch Instruction Simplified Mnemonics where bdnzt 4 cr5 eq target could be used instead of bdnzt 22 target This is describe...

Страница 453: ...Compare Word Simplified Mnemonics move to CR instructions and others can also modify CR fields so CR0 and CR1 may hold values that do not adhere to the meanings described in Table C 7 CR logical instr...

Страница 454: ...t 4 cr2 lt 4 cr3 lt 4 cr4 lt 4 cr5 lt 4 cr6 lt 4 cr7 lt 0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 000 001 010 011 100 101 110 111 00 Less than LT For integer compare instructions rA SIMM or rB sign...

Страница 455: ...cr4 100 48 51 cr5 101 52 55 cr6 110 56 59 cr7 111 60 63 Table C 10 Branch Simplified Mnemonics Branch Semantics LR Update Not Enabled LR Update Enabled bc bca bclr bcctr bcl bcla bclrl bcctrl Branch...

Страница 456: ...ror Subsequent examples test conditions 2 Same as 1 but branch only if CTR is nonzero and equal condition in CR0 bdnzt eq target equivalent to bc 8 2 target Other equivalents include bdnzt 2 target or...

Страница 457: ...ndition false bc 0 BI target bdnzf BI target bca 0 BI target bdnzfa BI target Decrement CTR branch if CTR 0 bc 18 0 target bdz target2 bca 18 0 target bdza target2 Decrement CTR branch if CTR 0 and co...

Страница 458: ...CTR 0 and condition true bcl 8 0 target bdnztl BI target bcla 8 BI target bdnztla BI target Decrement CTR branch if CTR 0 and condition false bcl 0 BI target bdnzfl BI target bcla 0 BI target bdnzfla...

Страница 459: ...R fields the test bit falls so the BI operand is replaced by a crS operand The standard codes shown in Table C 16 are used for the most common combinations of branch conditions Note that for ease of p...

Страница 460: ...crS 2 target_addr 2 BI can be a numeric value or an expression as shown in Table C 9 Branch Conditional to Link Register bclr bclrl BO BI bxlr bxlrl crS Branch Conditional to Count Register bcctr bcct...

Страница 461: ...c bca Simplified Mnemonic Branch if less than bc 12 BI 1 target 1 The value in the BI operand selects CRn 0 the LT bit blt crS target bca 12 BI1 target blta crS target Branch if less than or equal bc...

Страница 462: ...not equal bclr 4 BI3 target bnelr crS target bcctr 4 BI3 target bnectr crS target Branch if summary overflow bclr 12 BI 4 target 4 The value in the BI operand selects CRn 3 the SO bit bsolr crS targe...

Страница 463: ...n 1 the GT bit blelrl crS target bcctrl 4 BI2 target blectrl crS target Branch if not greater than bnglrl crS target bngctrl crS target Branch if equal bclrl 12 BI 3 target 3 The value in the BI opera...

Страница 464: ...as unsigned 32 bit integers and place result in CR0 cmplw rA rB equivalent to cmpl 0 0 rA rB C 6 Condition Register Logical Simplified Mnemonics The CR logical instructions shown in Table C 24 can be...

Страница 465: ...dopted for the most common combinations of trap conditions Table C 25 Standard Codes for Trap Instructions Code Description TO Encoding U 1 1 The symbol U indicates an unsigned less than evaluation is...

Страница 466: ...ctions evaluate a trap condition as follows The contents of rA are compared with either the sign extended SIMM field or the contents of rB depending on the trap instruction Table C 26 Trap Simplified...

Страница 467: ...destination GPR operand rS or rD Following are examples using the SPR simplified mnemonics 1 Copy the contents of rS to the XER mtxer rS equivalent to mtspr 1 rS 2 Copy the contents of the LR to rS m...

Страница 468: ...dis instructions can be used to load an immediate value into a register Additional mnemonics are provided to convey the idea that no addition is being performed but that data is being moved from the i...

Страница 469: ...g instruction copies the contents of rS into rA This mnemonic can be coded with a dot suffix to cause the Rc bit to be set in the underlying instruction mr rA rS equivalent to or rA rS rS C 9 5 Comple...

Страница 470: ...ent to isel rD rA rB 2 C 10 2 SPE Mnemonics The following mnemonic handles moving of the full 64 bit SPE GPR Vector Move evmr rD rA equivalent to evor rD rA rA The following mnemonic performs a comple...

Страница 471: ...crement CTR branch if CTR 0 bclrl with LR Update bdnzt BI target bc 8 BI target Decrement CTR branch if CTR 0 and condition true bc without LR update bdnzta BI target bca 8 BI target Decrement CTR bra...

Страница 472: ...10 BI target Decrement CTR branch if CTR 0 and condition true bcl with LR update bdztla BI target bcla 10 BI target Decrement CTR branch if CTR 0 and condition true bcla with LR update bdztlrl BI bclr...

Страница 473: ...updating bgela crS target bcla 4 BI4 target Branch if greater than or equal bcla with comparison conditions and LR updating bgelr crS target bclr 4 BI4 target Branch if greater than or equal bclr with...

Страница 474: ...blta crS target bca 12 BI4 target Branch if less than bca without comparison conditions or LR updating bltctr crS target bcctr 12 BI4 target Branch if less than bcctr without comparison conditions an...

Страница 475: ...t greater than bclr without comparison conditions and LR updating bnglrl crS target bclrl 4 BI5 target Branch if not greater than bclrl with comparison conditions and LR update bnl crS target bc 4 BI4...

Страница 476: ...target Branch if not unordered bcctrl with comparison conditions and LR update bnul crS target bcl 4 BI6 target Branch if not unordered bcl with comparison conditions and LR updating bnula crS target...

Страница 477: ...crS target bcctrl 12 BI6 target Branch if unordered bcctrl with comparison conditions and LR update bunl crS target bcl 12 BI6 target Branch if unordered bcl with comparison conditions and LR updating...

Страница 478: ...crf 0xFF rS Move to Condition Register mtspr rS mfspr SPRN rS Move to SPR see Section C 8 Simplified Mnemonics for Accessing SPRs nop ori 0 0 0 No op not rA rS nor rA rS rS NOT not rA rS nor rA rS rS...

Страница 479: ...twi 5 rA SIMM Trap immediate if logically not less than twlt rA SIMM tw 16 rA SIMM Trap if less than twlti rA SIMM twi 16 rA SIMM Trap immediate if less than twne rA SIMM tw 24 rA SIMM Trap if not equ...

Страница 480: ...PowerPC e500 Core Family Reference Manual Rev 1 C 36 Freescale Semiconductor Simplified Mnemonics for PowerPC Instructions...

Страница 481: ...1 1 1 1 1 rD rA rB 0 1 0 0 0 0 1 0 1 0 0 X add add 0 1 1 1 1 1 rD rA rB 0 1 0 0 0 0 1 0 1 0 1 X add addc 0 1 1 1 1 1 rD rA rB 0 0 0 0 0 0 1 0 1 0 0 X addc addc 0 1 1 1 1 1 rD rA rB 0 0 0 0 0 0 1 0 1 0...

Страница 482: ...0 0 0 1 0 0 1 1 0 0 X bblels bc 0 1 0 0 0 0 BO BI BD 0 0 B bc bca 0 1 0 0 0 0 BO BI BD 1 0 B bca bcctr 0 1 0 0 1 1 BO BI 1 0 0 0 0 1 0 0 0 0 0 XL bcctr bcctrl 0 1 0 0 1 1 BO BI 1 0 0 0 0 1 0 0 0 0 1 X...

Страница 483: ...uivalent to bca 2 BI target bdzfa bdzfl bdzfl BI target equivalent to bcl 2 BI target bdzfl bdzfla bdzfla BI target equivalent to bcla 2 BI target bdzfla bdzflr bdzflr BI equivalent to bclr 2 BI bdzfl...

Страница 484: ...cctrl 4 BI3 target bgectrl bgel bgel crS target equivalent to bcl 4 BI3 target bgel bgela bgela crS target equivalent to bcla 4 BI3 target bgela bgelr bgelr crS target equivalent to bclr 4 BI3 target...

Страница 485: ...t to bc 4 BI3 target bne bnea bnea crS target equivalent to bca 4 BI3 target bnea bnectr bnectr crS target equivalent to bcctr 4 BI3 target bnectr bnectrl bnectrl crS target equivalent to bcctrl 4 BI3...

Страница 486: ...rget equivalent to bca 4 BI5 target bnua bnuctr bnuctr crS target equivalent to bcctr 4 BI5 target bnuctr bnuctrl bnuctrl crS target equivalent to bcctrl 4 BI5 target bnuctrl bnul bnul crS target equi...

Страница 487: ...rA rS n n 32 equivalent to rlwinm rA rS 0 n 31 clrlwi clrrwi clrrwi rA rS n n 32 equivalent to rlwinm rA rS 0 0 31 n clrrwi cmp 0 1 1 1 1 1 crfD L rA rB 0 0 0 0 0 0 0 0 0 0 X cmp cmpi 0 0 1 0 1 1 crf...

Страница 488: ...1 1 0 1 1 0 X dcbz divw 0 1 1 1 1 1 rD rA rB 0 1 1 1 1 0 1 0 1 1 0 X divw divw 0 1 1 1 1 1 rD rA rB 0 1 1 1 1 0 1 0 1 1 1 X divw divwo 0 1 1 1 1 1 rD rA rB 1 1 1 1 1 0 1 0 1 1 0 X divwo divwo 0 1 1 1...

Страница 489: ...fD rA rB 0 1 0 1 1 1 1 1 1 0 1 EFX efdtstlt efsabs 0 0 0 1 0 0 rD rA 0 1 0 1 1 0 0 0 1 0 0 EFX efsabs efsadd 0 0 0 1 0 0 rD rA rB 0 1 0 1 1 0 0 0 0 0 0 EFX efsadd efscfd 0 0 0 1 0 0 rD 0 0 0 0 0 rB 0...

Страница 490: ...A rB 0 1 0 0 0 0 0 0 0 0 0 EVX evaddw evand 0 0 0 1 0 0 rD rA rB 0 1 0 0 0 0 1 0 0 0 1 EVX evand evandc 0 0 0 1 0 0 rD rA rB 0 1 0 0 0 0 1 0 0 1 0 EVX evandc evcmpeq 0 0 0 1 0 0 crfD rA rB 0 1 0 0 0 1...

Страница 491: ...g evfssub 0 0 0 1 0 0 rD rA rB 0 1 0 1 0 0 0 0 0 0 1 EVX evfssub evfststeq 0 0 0 1 0 0 crfD rA rB 0 1 0 1 0 0 1 1 1 1 0 EVX evfststeq evfststgt 0 0 0 1 0 0 crfD rA rB 0 1 0 1 0 0 1 1 1 0 0 EVX evfstst...

Страница 492: ...1 0 0 1 0 1 0 0 1 EVX evmhegsmiaa evmhegsmian 0 0 0 1 0 0 rD rA rB 1 0 1 1 0 1 0 1 0 0 1 EVX evmhegsmian evmhegumiaa 0 0 0 1 0 0 rD rA rB 1 0 1 0 0 1 0 1 0 0 0 EVX evmhegumiaa evmhegumian 0 0 0 1 0 0...

Страница 493: ...0 0 rD rA rB 1 0 1 1 0 0 0 1 1 1 1 EVX evmhosmfanw evmhosmi 0 0 0 1 0 0 rD rA rB 1 0 0 0 0 0 0 1 1 0 1 EVX evmhosmi evmhosmia 0 0 0 1 0 0 rD rA rB 1 0 0 0 0 1 0 1 1 0 1 EVX evmhosmia evmhosmiaaw 0 0...

Страница 494: ...rD rA rB 1 0 1 0 1 0 0 0 0 0 0 EVX evmwlusiaaw evmwlusianw 0 0 0 1 0 0 rD rA rB 1 0 1 1 1 0 0 0 0 0 0 EVX evmwlusianw evmwsmf 0 0 0 1 0 0 rD rA rB 1 0 0 0 1 0 1 1 0 1 1 EVX evmwsmf evmwsmfa 0 0 0 1 0...

Страница 495: ...wiu evsrws 0 0 0 1 0 0 rD rA rB 0 1 0 0 0 1 0 0 0 0 1 EVX evsrws evsrwu 0 0 0 1 0 0 rD rA rB 0 1 0 0 0 1 0 0 0 0 0 EVX evsrwu evstdd 0 0 0 1 0 0 rD rA UIMM 6 0 1 1 0 0 1 0 0 0 0 1 EVX evstdd evstddx 0...

Страница 496: ...1 0 0 1 1 0 1 0 1 X extsh icbi 0 1 1 1 1 1 rA rB 1 1 1 1 0 1 0 1 1 0 X icbi icblc 0 1 1 1 1 1 CT rA rB 0 0 1 1 1 0 0 1 1 0 0 X icblc icbt 0 1 1 1 1 1 CT rA rB 0 0 0 0 0 1 0 1 1 0 X icbt icbtls 0 1 1...

Страница 497: ...0 0 0 0 0 0 0 0 0 0 XL mcrf mcrxr 0 1 1 1 1 1 crfD 1 0 0 0 0 0 0 0 0 0 X mcrxr mfcr 0 1 1 1 1 1 rD 0 0 0 0 0 1 0 0 1 1 X mfcr mfcr mtcr rS equivalent to mtcrf 0xFF rS mfcr mfmsr 0 1 1 1 1 1 rD 0 0 0...

Страница 498: ...0 0 1 X nego nop nop equivalent to ori 0 0 0 nop nor 0 1 1 1 1 1 rS rA rB 0 0 0 1 1 1 1 1 0 0 0 X nor nor 0 1 1 1 1 1 rS rA rB 0 0 0 1 1 1 1 1 0 0 1 X nor not not rA rS equivalent to nor rA rS rS not...

Страница 499: ...m rA rS 32 n n 31 srwi stb 1 0 0 1 1 0 rS rA D D stb stbu 1 0 0 1 1 1 rS rA D D stbu stbux 0 1 1 1 1 1 rS rA rB 0 0 1 1 1 1 0 1 1 1 0 X stbux stbx 0 1 1 1 1 1 rS rA rB 0 0 1 1 0 1 0 1 1 1 0 X stbx sth...

Страница 500: ...subfo 0 1 1 1 1 1 rD rA rB 1 0 0 0 1 0 1 0 0 0 1 X subfo subfze 0 1 1 1 1 1 rD rA 0 0 1 1 0 0 1 0 0 0 0 X subfze subfze 0 1 1 1 1 1 rD rA 0 0 1 1 0 0 1 0 0 0 1 X subfze subfzeo 0 1 1 1 1 1 rD rA 1 0 1...

Страница 501: ...ent to twi 6 rA SIMM twlngi twlnl twlnl rA SIMM equivalent to tw 5 rA SIMM twlnl twlnli twlnli rA SIMM equivalent to twi 5 rA SIMM twlnli twlt twlt rA SIMM equivalent to tw 16 rA SIMM twlt twlti twlti...

Страница 502: ...0 1 0 1 1 0 0 0 0 0 0 EFX efsadd efscfsf 04 rD rB 0 1 0 1 1 0 1 0 0 1 1 EFX efscfsf efscfsi 04 rD rB 0 1 0 1 1 0 1 0 0 0 1 EFX efscfsi efscfuf 04 rD rB 0 1 0 1 1 0 1 0 0 1 0 EFX efscfuf efscfui 04 rD...

Страница 503: ...vcntlsw 04 rD rA 0 1 0 0 0 0 0 1 1 1 0 EVX evcntlsw evcntlzw 04 rD rA 0 1 0 0 0 0 0 1 1 0 1 EVX evcntlzw evdivws 04 rD rA rB 1 0 0 1 1 0 0 0 1 1 0 EVX evdivws evdivwu 04 rD rA rB 1 0 0 1 1 0 0 0 1 1 1...

Страница 504: ...evldh 04 rD rA UIMM 1 0 1 1 0 0 0 0 0 1 0 1 EVX evldh evldhx 04 rD rA rB 0 1 1 0 0 0 0 0 1 0 0 EVX evldhx evldw 04 rD rA UIMM 1 0 1 1 0 0 0 0 0 0 1 1 EVX evldw evldwx 04 rD rA rB 0 1 1 0 0 0 0 0 0 1...

Страница 505: ...B 1 0 1 0 0 0 0 1 0 1 1 EVX evmhesmfaaw evmhesmfanw 04 rD rA rB 1 0 1 1 0 0 0 1 0 1 1 EVX evmhesmfanw evmhesmi 04 rD rA rB 1 0 0 0 0 0 0 1 0 0 1 EVX evmhesmi evmhesmia 04 rD rA rB 1 0 0 0 0 1 0 1 0 0...

Страница 506: ...rD rA rB 1 0 0 0 0 1 0 0 1 1 1 EVX evmhossfa evmhossfaaw 04 rD rA rB 1 0 1 0 0 0 0 0 1 1 1 EVX evmhossfaaw evmhossfanw 04 rD rA rB 1 0 1 1 0 0 0 0 1 1 1 EVX evmhossfanw evmhossiaaw 04 rD rA rB 1 0 1 0...

Страница 507: ...0 1 EVX evmwsmi evmwsmia 04 rD rA rB 1 0 0 0 1 1 1 1 0 0 1 EVX evmwsmia evmwsmiaa 04 rD rA rB 1 0 1 0 1 0 1 1 0 0 1 EVX evmwsmiaa evmwsmian 04 rD rA rB 1 0 1 1 1 0 1 1 0 0 1 EVX evmwsmian evmwssf 04 r...

Страница 508: ...B 0 1 1 0 0 1 0 0 0 1 0 EVX evstdwx evstwhe 04 rS rA UIMM 3 0 1 1 0 0 1 1 0 0 0 1 EVX evstwhe evstwhex 04 rS rA rB 0 1 1 0 0 1 1 0 0 0 0 EVX evstwhex evstwho 04 rS rA UIMM 3 0 1 1 0 0 1 1 0 1 0 1 EVX...

Страница 509: ...1 0 0 0 0 1 XL bclrl crand 19 0x13 crbD crbA crbB 0 1 0 0 0 0 0 0 0 1 XL crand crandc 19 0x13 crbD crbA crbB 0 0 1 0 0 0 0 0 0 1 XL crandc creqv 19 0x13 crbD crbA crbB 0 1 0 0 1 0 0 0 0 1 XL creqv cr...

Страница 510: ...ddco adde 31 0x1F rD rA rB 0 0 1 0 0 0 1 0 1 0 0 X adde adde 31 0x1F rD rA rB 0 0 1 0 0 0 1 0 1 0 1 X adde addeo 31 0x1F rD rA rB 1 0 1 0 0 0 1 0 1 0 0 X addeo addeo 31 0x1F rD rA rB 1 0 1 0 0 0 1 0 1...

Страница 511: ...1 1 1 1 0 1 1 0 X dcbtst dcbtstls 31 0x1F CT rA rB 0 0 1 0 0 0 0 1 1 0 0 X dcbtstls dcbz 31 0x1F rA rB 1 1 1 1 1 1 0 1 1 0 X dcbz divw 31 0x1F rD rA rB 0 1 1 1 1 0 1 0 1 1 0 X divw divw 31 0x1F rD rA...

Страница 512: ...1 1 0 X mbar mcrxr 31 0x1F crfD 1 0 0 0 0 0 0 0 0 0 X mcrxr mfcr 31 0x1F rD 0 0 0 0 0 1 0 0 1 1 X mfcr mfmsr 31 0x1F rD 0 0 0 1 0 1 0 0 1 1 X mfmsr mfpmr 31 0x1F rD PMRN5 9 PMRN0 4 0 1 0 1 0 0 1 1 1...

Страница 513: ...1 1 0 0 0 0 X sraw sraw 31 0x1F rS rA rB 1 1 0 0 0 1 1 0 0 0 1 X sraw srawi 31 0x1F rS rA SH 1 1 0 0 1 1 1 0 0 0 0 X srawi srawi 31 0x1F rS rA SH 1 1 0 0 1 1 1 0 0 0 1 X srawi srw 31 0x1F rS rA rB 1 0...

Страница 514: ...x1F rD rA 0 0 1 1 0 0 1 0 0 0 0 X subfze subfze 31 0x1F rD rA 0 0 1 1 0 0 1 0 0 0 1 X subfze subfzeo 31 0x1F rD rA 1 0 1 1 0 0 1 0 0 0 0 X subfzeo subfzeo 31 0x1F rD rA 1 0 1 1 0 0 1 0 0 0 1 X subfzeo...

Страница 515: ...nemonic add 0 1 1 1 1 1 rD rA rB 0 1 0 0 0 0 1 0 1 0 0 X add add 0 1 1 1 1 1 rD rA rB 0 1 0 0 0 0 1 0 1 0 1 X add addc 0 1 1 1 1 1 rD rA rB 0 0 0 0 0 0 1 0 1 0 0 X addc addc 0 1 1 1 1 1 rD rA rB 0 0 0...

Страница 516: ...rS rA 0 0 0 0 0 1 1 0 1 0 1 X cntlzw dcba 0 1 1 1 1 1 rA rB 1 0 1 1 1 1 0 1 1 0 X dcba dcbf 0 1 1 1 1 1 rA rB 0 0 0 1 0 1 0 1 1 0 X dcbf dcbi 0 1 1 1 1 1 rA rB 0 1 1 1 0 1 0 1 1 0 X dcbi dcblc 0 1 1...

Страница 517: ...lhaux lhax 0 1 1 1 1 1 rD rA rB 0 1 0 1 0 1 0 1 1 1 X lhax lhbrx 0 1 1 1 1 1 rD rA rB 1 1 0 0 0 1 0 1 1 0 X lhbrx lhzux 0 1 1 1 1 1 rD rA rB 0 1 0 0 1 1 0 1 1 1 X lhzux lhzx 0 1 1 1 1 1 rD rA rB 0 1 0...

Страница 518: ...rA rB 0 1 1 0 0 1 1 1 0 0 1 X orc slw 0 1 1 1 1 1 rS rA rB 0 0 0 0 0 1 1 0 0 0 0 X slw slw 0 1 1 1 1 1 rS rA rB 0 0 0 0 0 1 1 0 0 0 1 X slw sraw 0 1 1 1 1 1 rS rA rB 1 1 0 0 0 1 1 0 0 0 0 X sraw sraw...

Страница 519: ...1 0 1 0 0 0 1 X subfo subfze 0 1 1 1 1 1 rD rA 0 0 1 1 0 0 1 0 0 0 0 X subfze subfze 0 1 1 1 1 1 rD rA 0 0 1 1 0 0 1 0 0 0 1 X subfze subfzeo 0 1 1 1 1 1 rD rA 1 0 1 1 0 0 1 0 0 0 0 X subfzeo subfzeo...

Страница 520: ...1 0 0 0 0 1 rD rA D D lwzu mulli 0 0 0 1 1 1 rD rA SIMM D mulli ori 0 1 1 0 0 0 rS rA UIMM D ori oris 0 1 1 0 0 1 rS rA UIMM D oris stb 1 0 0 1 1 0 rS rA D D stb stbu 1 0 0 1 1 1 rS rA D D stbu sth 1...

Страница 521: ...rB 0 1 0 1 1 1 0 1 0 0 1 EFX efddiv efdmul 0 0 0 1 0 0 rD rA rB 0 1 0 1 1 1 0 1 0 0 0 EFX efdmul efdnabs 0 0 0 1 0 0 rD rA 0 1 0 1 1 1 0 0 1 0 1 EFX efdnabs efdneg 0 0 0 1 0 0 rD rA 0 1 0 1 1 1 0 0 1...

Страница 522: ...0 0 EVX evabs evaddiw 0 0 0 1 0 0 rD UIMM rB 0 1 0 0 0 0 0 0 0 1 0 EVX evaddiw evaddsmiaaw 0 0 0 1 0 0 rD rA 1 0 0 1 1 0 0 1 0 0 1 EVX evaddsmiaaw evaddssiaaw 0 0 0 1 0 0 rD rA 1 0 0 1 1 0 0 0 0 0 1 E...

Страница 523: ...0 1 0 0 1 0 1 1 0 EVX evfsctuf evfsctui 0 0 0 1 0 0 rD rB 0 1 0 1 0 0 1 0 1 0 0 EVX evfsctui evfsctuiz 0 0 0 1 0 0 rD rB 0 1 0 1 0 0 1 1 0 0 0 EVX evfsctuiz evfsdiv 0 0 0 1 0 0 rD rA rB 0 1 0 1 0 0 0...

Страница 524: ...rA rB 0 1 0 0 0 1 0 1 1 0 0 EVX evmergehi evmergehilo 0 0 0 1 0 0 rD rA rB 0 1 0 0 0 1 0 1 1 1 0 EVX evmergehilo evmergelo 0 0 0 1 0 0 rD rA rB 0 1 0 0 0 1 0 1 1 0 1 EVX evmergelo evmergelohi 0 0 0 1...

Страница 525: ...D rA rB 1 0 1 1 0 1 0 1 1 0 1 EVX evmhogsmian evmhogumiaa 0 0 0 1 0 0 rD rA rB 1 0 1 0 0 1 0 1 1 0 0 EVX evmhogumiaa evmhogumian 0 0 0 1 0 0 rD rA rB 1 0 1 1 0 1 0 1 1 0 0 EVX evmhogumian evmhosmf 0 0...

Страница 526: ...rA rB 1 0 1 1 1 0 0 0 1 0 0 EVX evmwhusianw evmwlumi 0 0 0 1 0 0 rD rA rB 1 0 0 0 1 0 0 1 0 0 0 EVX evmwlumi evmwlumia 0 0 0 1 0 0 rD rA rB 1 0 0 0 1 1 0 1 0 0 0 EVX evmwlumia evmwlumiaaw 0 0 0 1 0 0...

Страница 527: ...1 1 0 EVX evslwi evsplatfi 0 0 0 1 0 0 rD SIMM 0 1 0 0 0 1 0 1 0 1 1 EVX evsplatfi evsplati 0 0 0 1 0 0 rD SIMM 0 1 0 0 0 1 0 1 0 0 1 EVX evsplati evsrwis 0 0 0 1 0 0 rD rA UIMM 0 1 0 0 0 1 0 0 0 1 1...

Страница 528: ...1 0 1 0 0 rS rA SH MB ME Rc M rlwimi rlwinm 0 1 0 1 0 1 rS rA SH MB ME 0 M rlwinm rlwinm 0 1 0 1 0 1 rS rA SH MB ME 1 M rlwinm rlwnm 0 1 0 1 1 1 rS rA rB MB ME Rc M rlwnm rlwnm 0 1 0 1 1 1 rS rA rB M...

Страница 529: ...0 0 0 0 0 1 XL crxor isync 0 1 0 0 1 1 0 0 1 0 0 1 0 1 1 0 XL isync mcrf 0 1 0 0 1 1 crfD crfS 0 0 0 0 0 0 0 0 0 0 XL mcrf rfci 0 1 0 0 1 1 0 0 0 0 1 1 0 0 1 1 XL rfci rfi 0 1 0 0 1 1 0 0 0 0 1 1 0 0...

Страница 530: ...Opcode Listings PowerPC e500 Core Family Reference Manual Rev 1 D 50 Freescale Semiconductor...

Страница 531: ...e e500v1 and e500v2 cores The coverage of Book E and Freescale Book E MMU architecture formerly in Chapter 13 Cache and MMU Background was removed See the EREF A reference for Freescale Book E and the...

Страница 532: ...ters Deleted MCSR GL_CI from Table 5 4 Also removed column Recoverable in same table Section 5 7 Interrupt Definitions Deleted references to ESR AP which is not implemented on the e500 Chapter 9 Timer...

Страница 533: ...aching allowed loads only if the two loads are to different 32 byte address granules with Newer non guarded caching allowed loads can bypass older non guarded caching allowed loads Chapter 12 Memory M...

Страница 534: ...PowerPC e500 Core Family Reference Manual Rev 1 E 4 Freescale Semiconductor Revision History...

Страница 535: ...tion select APU 3 25 3 60 machine check interrupt APU 3 63 5 2 see also Interrupt handling interrupt types machine check interrupt performance monitor APU 3 60 3 61 5 33 7 2 signal processing engine S...

Страница 536: ...quired bit M bit 11 12 global signaling M bit and snooping 11 12 instruction cache coherency model 11 8 11 11 address aliasing errors 11 8 maintaining in power down mode 6 3 see also Memory cache acce...

Страница 537: ...2 data address compare 8 9 instruction address compare 8 7 instruction complete debug event 8 12 interrupt taken debug event 8 13 return debug event 8 13 trap debug event 8 11 unconditional debug even...

Страница 538: ...12 2 12 12 12 20 12 22 12 23 12 24 trap instr exceptions program interrupt 5 6 5 12 5 24 Execution model self modifying code 3 17 Execution synchronization 3 11 Execution timing branch instructions 4...

Страница 539: ...d APUs embedded single precision floating point SPFP APUs 5 3 machine check interrupt APU 5 2 see also Auxiliary processing units APUs signal processing engine SPE APU 5 3 Freescale Book E implementat...

Страница 540: ...ngle precision floating point APUs 3 58 store serialization 4 16 system linkage 3 26 3 40 system register instruction latencies 4 31 TLB management instructions 3 41 12 17 12 24 synchronization requir...

Страница 541: ...idation protection 12 12 see also Memory management unit MMU TLBs isel instruction select APU 3 25 3 60 ISI instruction storage interrupt 5 20 see also Interrupt handling Issue stage see Execution tim...

Страница 542: ...nt algorithm true LRU 12 10 structure 12 9 L2 TLB arrays programmable 12 8 replacement algorithm general 12 13 replacement algorithm hints for round robin TLB0 12 13 structure 12 11 TLB0 4 Kbyte page...

Страница 543: ...ing for power savings 6 3 PLRU algorithm 11 25 see also Caches operation block replacement PMC0 3 performance monitor counter registers 2 57 7 8 PMGC0 global control register 0 1 31 2 53 7 4 PMLCa0 PM...

Страница 544: ...ntrol 0 PMGC0 7 4 global control register 0 PMGC0 1 31 2 53 local control A PMLCa0 PMLCa3 2 55 7 5 local control B PMLCb0 PMLCb3 2 56 7 6 PMR encodings 3 61 user counter registers UPMC0 3 2 58 7 9 use...

Страница 545: ...execution synchronization 3 11 general A 1 memory instructions 3 30 timing considerations 4 17 4 18 primitives A 2 compare and swap A 4 fetch and add A 3 fetch and AND A 3 fetch and no op A 2 fetch a...

Страница 546: ...47 UPMC0 3 user performance monitor counter registers 2 58 7 9 UPMGC0 user global control register 0 2 54 7 5 UPMLCa0 UPMLCa3 user performance monitor local control A registers 2 56 UPMLCa0 UPMLCa3 us...

Страница 547: ...Debug Support 8 Part II e500 Core Complex II Timer Facilities 9 Auxiliary Processing Units APUs 10 L1 Caches 11 Memory Management Units 12 Core Complex Bus CCB 13 Appendix A Programming Examples A Ap...

Страница 548: ...r 8 Debug Support II Part II e500 Core Complex 9 Timer Facilities 10 Auxiliary Processing Units APUs 11 L1 Caches 12 Memory Management Units 13 Core Complex Bus CCB A Appendix A Programming Examples B...

Результаты поиска

Содержание PowerPC e500 Core

Отзывы:

Похожие инструкции для PowerPC e500 Core

Бренды по названию

Популярные бренды

Freescale Semiconductor PowerPC e500 Core, Справочное руководство

Результаты поиска

Содержание PowerPC e500 Core

Отзывы:

Похожие инструкции для PowerPC e500 Core

Бренды по названию

Популярные бренды