| The Datasheet Archive - 100 Million Datasheets from 7500 Manufacturers. |
Proprietary Confidential Preliminary Issue February 2002 Propriet
Top Searches for this datasheetRM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary Proprietary Confidential Preliminary Issue February 2002 Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue ilic Data Sheet RM5261AMicroprocessor with 64-Bit System RM5261A RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue Document Information: document@pmc-sierra.com Corporate Information: info@pmc-sierra.com Technical Support: apps@pmc-sierra.com Site: http://www.pmc-sierra.com Tel: (604) 415-6000 Fax: (604) 415-6200 PMC-Sierra, Inc. 8555 Baxter Place Burnaby, Canada ilic Contacting PMC-Sierra RM5261A trademark PMC-Sierra, Inc. Trademarks event will PMC-Sierra, Inc. liable direct, indirect, special, incidental consequential damages, including, limited lost profits, lost business lost data resulting from reliance upon information, whether PMC-Sierra, Inc. been advised possibility such damage. None information contained this document constitutes express implied warranty PMCSierra, Inc. sufficiency, fitness suitability particular purpose such information fitness, suitability particular purpose, merchantability, performance, compatibility with other parts systems, products PMC-Sierra, Inc., portion thereof, referred this document. PMC-Sierra, Inc. expressly disclaims representations warranties kind regarding contents information, including, limited express implied warranties accuracy, completeness, merchantability, fitness particular use, non-infringement. Disclaimer PMC-2002240 (P3) information proprietary confidential PMC-Sierra, Inc., customers' internal use. event, cannot reproduce part this document, form, without express written consent PMC-Sierra, Inc. 2002 PMC-Sierra, Inc. rights reserved. Copyright Legal Information RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary Issue Date January 2002 Details Change Added RM5261A-400-H valid combinations. March 2001 Applied PMC-Sierra template existing (QED) FrameMaker document. Revised features list, Absolute Maximum Ratings table, Recommended Operating Conditions table, Electrical Characteristics table, Power Consumption table, Clock Parameters table System Interface Parameters table. Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue ilic Modified SysClock Frequency SysClock Period values Clock Parameters table. Added reference VccInt Power Consumption table. Changed standby modes 350. Changed maximum worst case instruction 1250. Modified Note August 2001 Added feature: 1.65 core with (p9). Changed recommended operating conditions VccInt 1.57 1.85 VccP 1.57 1.85 Added VssP commercial industrial values. Modified Note Modified recommended operating conditions table (page 31), power consumption table (page 33), clock parameters table (page 34), system interface parameters table (page include values operation. Modified system interface description (page read peak rate GByte/ with SysClock. Modified features include operating frequency, Dhrystone MIPS, MFLOPS. Issue Revision History RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue ilic signal, pin, names described text, such ExtRqst*, boldface typeface. field names described text, such Interrupt Mask, italic-bold typeface. instruction names, such MFHI, serif typeface. following conventions used this datasheet: Document Conventions RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary Legal Information Revision History Table Contents List Figures Document Conventions Integer Unit Pipeline 3.10 Floating-Point General Register File 3.12 System Control Co-Processor Registers 3.13 Virtual Physical Address Mapping 3.14 Joint 3.16 Data 3.18 Instruction Cache 3.20 Write buffer 3.21 System Interface 3.22 System Address/Data 3.23 System Command Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue 3.24 Handshake Signals 3.25 Non-overlapping System Interface 3.26 Enhanced Write Modes 3.27 External Requests 3.28 Interrupt Handling 3.29 Standby Mode 3.30 JTAG Interface 3.19 Data Cache 3.17 Cache Memory ilic 3.15 Instruction 3.11 System Control Co-processor (CP0) Floating-Point Unit Floating-Point Co-Processor Integer Multiply/Divide Register File Registers Superscalar Dispatch Hardware Overview Block Diagram Features List Tables Table Contents RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary 3.32 Boot-Time Modes Absolute Maximum Ratings Power Consumption Electrical Characteristics Capacitive Load Deration System Interface Parameters Boot-Time Interface Parameters Timing Diagrams Packaging Information RM5261A 208-QFP Package Alphabetical Pinout Ordering Information Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue ilic RM5261A 208-QFP Package Numerical Pinout 10.1 System Interface Timing (SysAD, SysCmd, ValidIn*, ValidOut*, etc.) Clock Parameters Electrical Characteristics Recommended Operating Conditions Descriptions 3.31 Boot-Time Options RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary Figure Block Diagram Figure Registers Figure Registers Figure Kernel Mode Virtual Addressing (32-bit) Figure Pipeline Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue ilic Figure Output Timing Figure Input Timing Figure Clock Timing Figure Processor Block Write Figure Processor Block Read Figure Typical Embedded System Block Diagram List Figures RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary Table Integer Multiply/Divide Operations Table Floating-Point Instruction Cycles Table Boot-Time Mode Stream Table System Interface Table Cache Attributes Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue ilic Table Power Supply Table Initialization Interface Table JTAG Interface Table Interrupt Interface Table Clock/Control Interface List Tables RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue ilic High-performance floating-point unit: MFLOPS Single cycle repeat rate common single-precision operations some double-precision operations cycle repeat rate double-precision multiply double precision combined multiply-add operations Single cycle repeat rate single-precision combined multiply-add operation MIPS instruction Floating point multiply-add instruction increases performance signal processing graphics applications Conditional moves reduce branch frequency Index address modes (register register) Embedded application enhancements Specialized integer Multiply-Accumulate instructions 3-operand multiply instruction cache locking Optional dedicated exception vector interrupts Fully static 0.18 micron CMOS design with power down logic Standby reduced power mode with WAIT instruction 1.65 core with 208-pin package Dual Issue superscalar microprocessor 250, 300, 350, operating frequencies Dhrystone MIPS High-performance system interface 64-bit multiplexed system address/data optimum price/performance High-performance write protocols maximize uncached write bandwidth Processor clock multipliers 2.5, 3.5, 4.5, IEEE 1149.1 JTAG boundary scan Integrated on-chip caches instruction data associative locking Virtually indexed, physically tagged Write-back write-through page basis Pipeline restart first doubleword data cache misses Integrated memory management unit Fully associative joint (shared translations) dual entries pages Variable page size increments) Features RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary DTLB ITLB Store Buffer Write Buffer Read Buffer Buffer Address Buffer Instruction Dispatch Unit Integer Instruction Register Instruction Register Floating-Point Control Floating-Point Load/Align Floating-Point Register File Packer/Unpacker Joint Coprocessor System/Memory Control Load Aligner Integer Control Integer Register File Integer Address/Adder Shifter/Store Aligner Logic Unit Branch Adder ITLB Virtual DTLB Virtual PLL/Clocks Mult, Div, Madd Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue ilic Program Counter Floating-Point MultAdd, Add, Sub, Cvt, Div, Sqrt Incrementer Primary Data Cache 2-way Associative DTag ITag Primary Instruction Cache 2-way Associative Integer Figure Block Diagram Block Diagram RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary General Purpose Registers Multiply/Divide Registers ilic Program Counter Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue Integer Unit RM5261A implements MIPS Instruction Architecture therefore fully upward compatible with applications that processors implementing earlier generation MIPS IIII instruction sets. Additionally, RM5261A includes implementation specific instructions found baseline MIPS that useful embedded market place. These instructions integer multiply-accumulate (MAD) 3-operand integer multiply (MUL). Figure Registers RM5261A contains general purpose registers, special purpose registers integer multiplication division, program counter, condition code bits. Figure shows user visible state. Registers RM5261A asymmetric superscalar dispatch unit which allows issue integer instruction floating-point computation instruction simultaneously. Integer instructions include alu, branch, load/store, floating-point load/store, while floating-point computation instructions include floating-point add, subtract, combined multiply-add, convert. combination with high-throughput fully pipelined floating-point execution unit, superscalar capability RM5261A provides unparalleled price/performance computationally intensive embedded applications. Superscalar Dispatch RM5261A offers high-level integration targeted high-performance embedded applications. elements RM5261A briefly described below. Hardware Overview RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary Figure Pipeline Figure shows RM5261A integer pipeline. illustrated figure, five integer instructions executing simultaneously. RM5261A multiplies input SysClock 2.5, 3.5, 4.5, produce pipeline clock. cycle Register File Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue RM5261A consists integer adder/subtractor, logic unit, shifter. adder performs address calculations addition arithmetic operations. logic unit performs logical zero shift data moves. shifter performs shifts store alignment operations. Each these units optimized perform operations single processor cycle. RM5261A thirty-two general purpose registers with register location (r0) hard-wired zero value. These registers used scalar integer operations address calculation. register file read ports write port fully bypassed minimize operation latency pipeline. 1I-1R: Instruction cache access Instruction virtual physical address translation Register file read, Bypass calculation, Instruction decode, Branch address calculation Issue slip decision, Branch decision Data virtual address calculation 1A-2A: Integer add, logical, shift Store Align 2A-2D: Data cache access load align Data virtual physical address translation Register file write ilic integer operations, loads, stores, other non-floating-point operations, RM5261A implements 5-stage integer pipeline. addition integer pipeline, RM5261A implements extended 7-stage pipeline floating-point operations. Pipeline RM5261A integer unit includes thirty-two general purpose 64-bit registers, load/store architecture with single cycle operations (add, sub, logical, shift) autonomous multiply/divide unit. Additional register resources include: HI/LO result registers twooperand integer multiply/divide operations, program counter (PC). RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary DMULT, DMULTU DIV, DIVD DDIV, DDIVU Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue RM5261A incorporates high-performance fully pipelined floating-point co-processor which includes floating-point register file autonomous execution units multiply/add/convert divide/square root. floating-point coprocessor tightly coupled execution unit, decoding executing instructions parallel with, case floating-point loads stores, cooperation with integer unit. superscalar capabilities RM5261A allow floatingpoint computation instructions issue concurrently with integer instructions. Floating-Point Unit RM5261A floating-point execution unit supports single double precision arithmetic, specified IEEE Standard 754. execution unit broken into separate divide/square root Floating-Point Co-Processor multiply-add instructions, MADU, multiply operands resulting product current contents registers. multiply-accumulate operation core primitive almost signal processing algorithms, allowing RM5261A eliminate need separate engine many embedded applications. ilic addition baseline MIPS integer multiply instructions, RM5261A also implements 3-operand multiply instruction, MUL. This instruction specifies that multiply result directly integer register file rather than register. portion multiply that would have normally gone into register discarded. applications where known that upper half multiply result required, using instruction eliminates necessity executing explicit MFLO instruction. baseline MIPS specifies that results multiply divide operation placed registers. These values then transferred general purpose register file using Move-from-Hi Move-from-Lo (MFHI/MFLO) instructions. MULT/U, MAD/U Opcode Operand Size Latency Repeat Rate Stall Cycles Table Integer Multiply/Divide Operations RM5261A dedicated integer multiply/divide unit optimized high-speed multiply multiply-accumulate operations. Table shows performance multiply/divide unit each operation. Integer Multiply/Divide RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary Floating-point operations include: subtract multiply divide square root reciprocal reciprocal square root conditional moves conversion between fixed-point floating-point format conversion between floating-point formats floating-point compare Table gives latencies floating-point instructions internal processor cycles. Table Floating-Point Instruction Cycles fadd fsub fmult fmadd fmsub fdiv fsqrt frecip frsqrt ilic 21/36 21/36 21/36 38/68 fcvt.s.d fcvt.s.w fcvt.s.l fcvt.d.s fcvt.d.w Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue Operation Latency Repeat Rate 19/34 19/34 19/34 36/66 RM5261A maintains fully precise floating-point exceptions while allowing both overlapped pipelined operations. Precise exceptions extremely important object-oriented programming environments highly desirable debugging environment. unit pipelined multiply/add unit. Overlap divide/square root multiply/add operations supported. RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary fcvt.d.l fcvt.w.s fcvt.w.d fcvt.l.s fcvt.l.d fcmp fmov fmovc fabs fneg Numbers represented single/double precision format. Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue memory management unit controls virtual memory system page mapping. consists instruction address translation buffer, ITLB, data address translation buffer, DTLB, Joint instruction data address translation buffer, JTLB, co-processor registers used virtual memory mapping sub-system. system control coprocessor, also called coprocessor MIPS architecture, responsible virtual memory sub-system, exception control system, diagnostics capability processor. MIPS architecture, system control co-processor (and thus kernel software) implementation dependent. 3.11 System Control Co-processor (CP0) floating-point control register space contains registers; determining configuration revision information coprocessor, control status information. These primarily used diagnostic software, exception handling, state saving restoring, control rounding modes. support superscalar operation, four read ports write ports, fully bypassed minimize operation latency pipeline. Three read ports write port used support combined multiply-add instruction while fourth read second write port allows concurrent floating-point load store. ilic floating-point general register file (FGR) made thirty-two 64-bit registers. With floating-point load store double instructions (LDC1 SDC1) floating-point unit take advantage 64-bit wide data cache issue floating-point coprocessor load store doubleword instruction every cycle. 3.10 Floating-Point General Register File Notes: Operation Latency Repeat Rate RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary RM5261A incorporates system control co-processor (CP0) registers on-chip. These registers provide path through which virtual memory system's page mapping examined modified, exceptions handled, operating modes controlled (kernel user mode, interrupts enabled disabled, cache features). addition, RM5261A includes registers implement real-time cycle counting facility cache diagnostic testing assist data error detection. Figure Registers PageMask EntryHi EntryLo0 EntryLo1 Figure shows registers. Context Count Status BadVAddr Compare Cause Random Wired PRId Config Index (entries protected from TLBWR) LLAddr TagLo ilic TagHi Register number 3.13 Virtual Physical Address Mapping user mode kernel mode supervisor mode Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue This mechanism allows system software provide secure environment user processes. Bits register Status determine which virtual addressing mode used. user mode, RM5261A provides single, uniform virtual address space 32-bit mode). When operating kernel mode, four distinct virtual address spaces, totalling over 32-bit mode), simultaneously available differentiated high-order bits virtual address. RM5261A provides three modes virtual addressing: Used memory management XContext CacheErr ErrorEPC Used exception processing 3.12 System Control Co-Processor Registers RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary Figure shows address space layout 32-bit operation Figure Kernel Mode Virtual Addressing (32-bit) 0xFFFFFFFF Kernel virtual address space (kseg3) 0xE0000000 Mapped, 0.5GB (ksseg) 0xC0000000 Mapped, 0.5GB 0xA0000000 0x9FFFFFFF Cached kernel physical address space (kseg0) 0x80000000 Unmapped, 0.5GB 3.14 Joint Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue fast virtual-to-physical address translation, RM5261A uses large, fully associative that maps virtual pages their corresponding physical addresses. indicated name, joint (JTLB) used both instruction data translations. JTLB organized pairs even-odd entries, maps virtual address address space identifier into large, physical address space. mechanisms provided assist controlling amount mapped space replacement characteristics various memory regions. First, page size configured, per-entry basis, page sizes range multiples Page Mask register loaded with desired page size mapping, that size stored into 0x00000000 ilic Mapped, 2.0GB (kuseg) 0x7FFFFFFF User virtual address space Unmapped, 0.5GB (kseg1) 0xBFFFFFFF Uncached kernel physical address space 0xDFFFFFFF Supervisor virtual address space When RM5261A configured 64-bit microprocessor, virtual address space layout upward compatible extension 32-bit virtual address space layout. RM5261A processors also support supervisor mode which virtual address space over (2.5 32-bit mode), divided into three regions based high-order bits virtual address. RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary 3.16 Data Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue RM5261A implements 4-entry data (DTLB) same reasons cited above ITLB. Each DTLB entry maps page. DTLB improves performance allowing data address translation occur parallel with instruction address translation. When miss occurs data address translation DTLB, DTLB filled from JTLB. DTLB refill pseudo-LRU: least recently used entry least recently used pair entries filled. operation DTLB completely transparent user. RM5261A implements 2-entry instruction (ITLB) minimize contention JTLB, eliminate timing critical path translating through large associative array, save power. Each ITLB entry maps page. ITLB improves performance allowing instruction address translation occur parallel with data address translation. When miss occurs instruction address translation ITLB, least-recently used ITLB entry filled from JTLB. operation ITLB completely transparent user. 3.15 Instruction ilic coherency attributes generate coherent transaction types system interface. However, RM5261A cache coherency supported. Hence coherency attributes should never used. non-coherent protocols used both code data RM5261A, with data using write-back write-through depending application. uncached non-coherent write-back non-coherent write-through with write-allocate non-coherent write-through without write-allocate sharable exclusive update JTLB also contains information that controls cache coherency protocol each page. Specifically, each page attribute bits determine whether coherency algorithm following: second mechanism controls replacement algorithm when miss occurs. RM5261A provides random replacement algorithm select entry written with mapping; however, processor also provides mechanism whereby system specific number mappings locked into TLB, thereby avoiding random replacement. This mechanism uses Wired register allows operating system guarantee that certain pages always mapped performance reasons deadlock avoidance. This mechanism also facilitates design real-time systems allowing deterministic access critical software. along with virtual address when entry written. Thus, operating systems create special purpose maps; example, entire frame buffer memory mapped using only entry. RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary 3.18 Instruction Cache Cache miss refill writes bits cycle minimize cache miss penalty. line size eight instructions bytes) maximize performance communication between processor memory system. RM5261A supports cache locking. contents cache locked setting coprocessor Status register. Locking prevents contents from being overwritten subsequent cache miss. Refills occur only into This mechanism allows programmer lock critical code into cache, thereby guaranteeing deterministic behavior locked code sequence. Uncached Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue Data loads instruction fetches from uncached memory space brought from main memory register file execution unit, respectfully. caches accessed. Data stores uncached memory space directly main memory without updating data cache. Write-back Loads instruction fetches first search cache, reading main memory only desired data cache resident. data store operations, cache first searched determine Cache protocols supported data cache are: data cache protected with byte parity protected with single parity bit. virtually indexed physically tagged allow simultaneous address translation data cache access. fast, single cycle data access, RM5261A includes on-chip data cache that twoway associative with fixed 32-byte (eight words) line size. ilic 3.19 Data Cache instruction cache 64-bits wide accessed each processor cycle. Accessing bits cycle allows instruction cache supply instructions cycle superscalar dispatch unit. typical code sequences where floating-point load store floating-point computation instruction being issued together loop, entire bandwidth available from instruction cache consumed. Since cache virtually indexed, virtual-to-physical address translation occurs parallel with cache access, further increasing performance allowing these operations occur simultaneously. cache contains 24-bit physical address, valid bit, single parity bit. RM5261A incorporates two-way associative on-chip instruction cache. This virtually indexed, physically tagged cache size protected with word parity. RM5261A incorporates on-chip instruction data caches that accessed single processor cycle. Each cache 64-bit data path both caches accessed simultaneously. cache subsystem provides integer floating-point units with aggregate bandwidth second internal clock frequency MHz. 3.17 Cache Memory RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary most commonly used write policy write-back, where store cache line does immediately cause main memory updated. This increases system performance reducing traffic eliminating bottleneck waiting each store operation finish before issuing subsequent memory operation. Software can, however, select write-through perpage basis when appropriate, such frame buffers. Associated with data cache store buffer. When RM5261A executes store instruction, this single-entry buffer gets written with store data while comparison performed. matches, then data written into data cache next cycle that data cache accessed (the next non-load cycle). store buffer allows RM5261A execute store every processor cycle perform back-to-back stores without penalty. event store immediately followed load same address, combined merge cache write occurs such that penalty incurred. RM5261A cache attributes both instruction data caches summarized Table Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue ilic Loads instruction fetches first search cache, reading main memory only desired data cache resident. data store operations, cache first searched determine target address cache resident. resident, cache contents updated main memory written, leaving write-back cache line unchanged. cache lookup misses, then only main memory written. Write-through without write allocate Loads instruction fetches first search cache, reading main memory only desired data cache resident. data store operations, cache first searched determine target address cache resident. resident, cache contents updated main memory written, leaving write-back cache line unchanged. cache lookup misses, target line first brought into cache then write performed above. Write-through with write allocate target address cache resident. resident, cache contents updated, cache line marked later write-back. cache lookup misses, target cache line first brought into cache then write performed above. RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary Characteristics Size Organization Line size Index Write policy Read order Write order miss restart after transfer Parity Cache locking Instruction 32KB 2-way associative vAddr11.0 pAddr31.12 n.a. sub-block sequential entire line per-word Data 32KB Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue system interface consists 64-bit Address/Data with parity check bits 9-bit command bus. addition, there handshake signals interrupt inputs. interface capable transferring data between processor memory peak rate GB/sec with SysClock. ilic 3.21 System Interface Writes external memory, whether cache miss write-backs stores uncached write-through addresses, on-chip write buffer. write buffer holds four 64-bit address data pairs. entire buffer used data cache write-back allows processor proceed parallel with memory update. uncached write-through stores, write buffer significantly increases performance decoupling SysAD transfers from instruction execution stream. 3.20 Write buffer sub-block sequential first double per-byte write-back/write-through pAddr31.12 vAddr11.0 2-way associative Table Cache Attributes RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary Control DRAM Address Flash/ Boot RM5261A Memory Controller Latch 3.23 System Command RM5261A interface 9-bit System Command (SysCmd) bus. command indicates whether SysAD carries address data information per-clock basis. SysAD carries address, SysCmd indicates what type transaction take place (for example, read write). SysAD carries data, SysCmd provides information about data (for example, this last data word transmitted, data contains error). SysCmd bidirectional support both processor requests external requests RM5261A. Processor requests initiated RM5261A responded external device. External requests issued external device require RM5261A respond. 3.24 Handshake Signals Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue RM5261A supports one- eight-byte transfers well block transfers SysAD bus. case sub-double word transfer, three low-order address bits give byte address transfer, SysCmd indicates number bytes being transferred. There handshake signals system interface. these, RdRdy* WrRdy*, used external device indicate RM5261A whether accept read write ilic Block Write data rate, Non-block Write protocol, Output Drive strength programmable Boot time Mode Control bits. rate which processor receives data also fully controlled external device. 64-bit System Address Data (SysAD) used transfer addresses data between RM5261A rest system. protected with 8-bit parity check (SysADC). system interface configurable allow easy interfacing memory systems varying frequencies. 3.22 System Address/Data Figure Typical Embedded System Block Diagram Figure shows typical embedded system using RM5261A. this example, bank DRAMs memory controller ASIC share processor's SysAD while memory controller provides separate ports boot system. RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary SysCmd ValidOut* ValidIn* Read SysAD Addr ilic SysClock Figure Processor Block Read Figure shows processor block read request external agent read response. read latency cycles (ValidOut* ValidIn*), response data pattern DDDD, indicating that data transferred every clock with wait states in-between. Data0 NData processor reads RM5261A asserts ValidOut* simultaneously drives address read command SysAD SysCmd buses respectively. system interface RdRdy* asserted, then processor tristates drivers releases system interface slave state asserting Release*. external device then begin sending data RM5261A. Data1 NData Data2 NData RdRdy* WrRdy* Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue Release* RM5261A implements non-overlapping system interface, meaning that only processor request outstanding time that request must serviced external device before RM5261A issues another request. RM5261A issue read write requests external device, whereas external device issue null write requests RM5261A. Data3 NEOD 3.25 Non-overlapping System Interface ValidOut* ValidIn* used RM5261A external device respectively indicate that there valid address, command, data SysAD SysCmd buses. RM5261A asserts ValidOut* when driving these buses with valid address, command, data, external device drives ValidIn* when control buses driving valid address, command, data. ExtRqst* Release* used transfer control SysAD SysCmd buses from processor external device. When external device needs control interface, asserts ExtRqst*. RM5261A responds asserting Release* release system interface slave state. transaction. RM5261A samples these signals before deasserting address read write requests. RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary Figure SysClock Processor Block Write ValidOut* ValidIn* WrRdy* Release* External Request pin, ExtRqst*, asserted external agent when requires mastership system interface, either perform independent transfer write interrupt register within RM5261A. independent transfer data transfer between external agents between external agent memory peripheral system interface. Following asserting ExtRqst*, RM5261A tri-states drivers allowing external agent system interface buses complete independent transfer. external agent responsible returning mastership system interface RM5261A when completed independent transfer does executing External Null cycle. Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue 3.27 External Requests write reissue mode, write rate write every cycles achieved. Pipelined writes have same cycle write repeat rate, issue additional write following deassertion WrRdy*. Write reissue mode enhancement pipelined write mode allows processor reissue missed write cycles. WrRdy* deasserted during issue phase write operation, cycle aborted processor reissued later time. ilic Pipelined write mode eliminates these wait states allowing processor drive write address onto immediately after previous write data cycle. This allows higher SysAD utilization. However, high frequencies processor drive subsequent write onto prior time external agent deasserts WrRdy*, indicating that accept another write cycle. This cause write cycle missed. RM5261A implements enhancements original R4000 write mechanism: Write Reissue Pipeline Writes. original R4000 allowed write address cycle SysAD only once every four SysClock cycles. Hence non-block write, this meant that every four cycles were wait states. 3.26 Enhanced Write Modes RdRdy* SysCmd Write NData NData NData NEOD SysAD Addr Data0 Data1 Data2 Data3 Figure shows processor block write using write response pattern DDDD, code boottime mode select options. RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary 3.29 Standby Mode 3.30 JTAG Interface 3.31 Boot-Time Options Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue boot-time serial mode stream defined Table presented processor first stream when VCCOK asserted. last transferred. 3.32 Boot-Time Modes Immediately after VCCOK signal asserted, processor reads serial stream bits initialize fundamental operational modes. ModeClock runs continuously from assertion VCCOK. Fundamental operational modes processor initialized boot-time mode control interface. This serial interface operates very frequency (SysClock divided 256). frequency operation allows initialization information kept cost EPROM system interface ASIC. ilic RM5261A interface supports JTAG Test Access Port (TAP) boundary scan conformance with IEEE 1149.1 specification. JTAG interface especially helpful checking integrity processors connections. Executing WAIT instruction enables interrupts causes processor enter Standby Mode. SysAD idle when wait instruction completes pipe stage, internal processor clock stops pipeline suspended. phase lock loop, PLL, internal timer/ counter, "wake input pins: Int[5:0]*, NMI*, ExtReq*, Reset*, ColdReset* continue operate their normal fashion. SysAD idle when WAIT instruction completes pipe-stage, then WAIT treated until operation completed. Once processor Standby, interrupt, including internally generated timer interrupt, causes processor exit Standby mode resume operation where left off. WAIT instruction typically inserted idle loop operating system real time executive. RM5261A provides means reduce amount power consumed internal core when performing useful operations. This state known Standby Mode. RM5261A supports dedicated interrupt vector. When enabled real time executive setting Cause register), interrupts vector specific address that shared with other exception types. This capability eliminates need through normal software routine exception decode dispatch, thereby lowering interrupt latency. 3.28 Interrupt Handling RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary Pclock SysClock Multiplier Mode Bits Mode 20=0 Multiply Multiply Multiply Multiply Multiply Multiply Multiply Multiply 19:18 DDDD DDxDDx DDxxDDxx DxDxDxDx DDxxxDDxxx DDxxxxDDxxxx DxxDxxDxxDxx DDxxxxxxDDxxxxxx DxxxDxxxDxxxDxxx 9-15 reserved 10:9 Non-Block Write Protocol R4000 compatible reserved pipelined write re-issue ilic Little endian endian Specifies byte ordering. Logically ORed with BigEndian input signal. Multiply Multiply Multiply Mode 20=1 Timer Interrupt Enable/Disable Enable timer interrupt Int5* Disable timer interrupt Int5* Reserved: Must zero Output driver strength 100% fastest strength strength 100% strength strength 255:23 Reserved: Must zero 14:13 Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue Reserved: Must zero Select Pclock SysClock Multiply Mode Integer Multipliers Half-Integer Multipliers External Width 64-bit 32-bit VccIO Setting VccIO 3.3V VccIO 2.5V Write-back data rate 17:16 System configuration identifiers software visible Config[21.20] register reserved (must zero) Reserved: Must zero Mode Description Mode Description Table Boot-Time Mode Stream RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary Name ExtRqst* Type Input Description External Request RdRdy* Input Read Ready ValidOut* Output Valid Output SysAD[63:0] Input/Output ilic SysADC[7:0] Input/Output SysCmd[8:0] Input/Output SysCmdP Input/Output Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue Signals that processor driving valid address data SysAD valid command data identifier SysCmd bus. System Address/Data 64-bit address data communication between processor external agent. System Address/Data check 8-bit containing parity check bits SysAD during data cycles. System Command/Data identifier 9-bit command data identifier transmission between processor external agent. Reserved system Command/Data identifier parity RM5261A, unused input zero output. Signals that external agent driving valid address data SysAD valid command data identifier SysCmd bus. ValidIn* Input Valid Input Signals that external agent accept processor write request. WrRdy* Input Write Ready Signals that external agent accept processor read. Signals that processor releasing system interface slave state. Release* Output Release Interface Signals that system interface submitting external request. Table System Interface following list interface, interrupt, miscellaneous pins available RM5261A. asterisk signal name denotes active-low. Descriptions RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary Name SysClock Type Input Description System Clock Master clock input used system interface reference clock. output timings relative this input clock. Pipeline operation frequency derived multiplying this clock factor selected during boot initialization. Int[5:0]* Input Interrupt NMI* Input Non-maskable interrupt JTDI Input ilic Name Type JTCK Input JTDO Output JTMS Input Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue Table JTAG Interface Description JTAG data JTAG serial data JTAG clock input JTAG serial clock input. JTAG data JTAG serial data out. JTAG command JTAG command signal, signals that incoming serial data command data. Non-maskable interrupt, ORed with interrupt register. general processor interrupts, bit-wise ORed with bits interrupt register. Name Type Description Table Interrupt Interface Quiet internal phase locked loop. Must connected through filter circuit. VSSP Input Quiet Quiet internal phase locked loop. Must connected VccInt through filter circuit. VccP Input Quiet Table Clock/Control Interface RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary Name BigEndian VCCOK Type Input Input Description Reset* Input Reset ModeClock Output Boot mode clock VCCIO Input VCCInt Input ilic Name Type Input Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue Table Power Supply Description Power supply core. Power supply I/O. Ground return. Serial boot-mode data input. ModeIn Input Boot mode data Serial boot-mode data clock output system clock frequency divided 256. This signal must asserted reset sequence. asserted synchronously asynchronously cold reset, synchronously initiate warm reset. Reset must de-asserted synchronously with SysClock. This signal must asserted power reset cold reset. ColdReset must de-asserted synchronously with SysClock. ColdReset* Input Cold reset When asserted, this signal indicates RM5261A that both power supplies been above recommended value more than milliseconds will remain stable. assertion VCCOK initiates reading boot-time mode control serial stream. Allows system change processor addressing mode without rewriting mode ROM. Table Initialization Interface RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary Symbol VTERM TCASE Rating Terminal Voltage with respect Operating Temperature Commercial Industrial Limits -0.52 +3.9 IOUT Input Current Output Current ±203 ±204 TSTG Storage Temperature +125 Notes: When VCCIO more than output should shorted time. Duration short should exceed seconds. Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue ilic Stresses greater than those listed under ABSOLUTE MAXIMUM RATINGS cause permanent damage device. This stress rating only functional operation device these other conditions above those indicated operational sections this specification implied. Exposure absolute maximum rating conditions extended periods affect reliability. minimum -2.0 pulse width less than should exceed Volts. Unit Absolute Maximum Ratings1 RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary Grade Commercial Speed VCCIO Case Temperature VCCInt +85°C +70°C VCCP VSSP 3.15 3.45 Recommended Operating Conditions 1.57 1.85 3.15 3.45 1.57 1.85 Industrial -40°C +85°C Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue ilic specified IEEE 1149.1 (JTAG), JTMS must held high during reset avoid entering JTAG test mode. VCCP must connected VCCInt through passive filter circuit. VSSP must connected through passive filter circuit. RM5200 User's Manual recommended filter circuit. Applying logic high state before VCCInt becomes stable recommended. VCCIO should exceed VCCInt greater than during power-up sequence. Notes: 1.57 1.85 3.15 3.45 1.57 1.85 RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary VccIO -0.3 VccIO VCCIO Parameter Minimum Maximum |IOUT| VccIO |IOUT|= |IOUT| -0.3 ilic |IOUT| VccIO VccIO Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue |IOUT|= Conditions Parameter Minimum Maximum Conditions VCCIO 3.15 3.45 Electrical Characteristics RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue ilic Dhrystone instruction mix. VCCIO supply power application dependant, typically <20% VCCInt. Maximum supply voltage (VCCInt) with maximum temperature (TCase). Notes: Maximum worst case instruction 1250 1400 VCCInt Power (mWatts)3 standby active Maximum with operation2 1150 1350 Parameter Conditions Max1 Max1 Max1 Max1 1800 1450 1600 Speed 2100 Power Consumption RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary Speed Test Parameter SysClock High SysClock SysClock Frequency1 SysClock Period Clock Jitter SysClock SysClock Rise Time SysClock Fall Time tSCP tSCH tSCL Transition Transition Symbol Conditions Clock Parameters Load Derate Parameter Symbol Capacitive Load Deration Units ±150 ±150 ±150 Units ns/25pF ±150 tSCP tSCP Electrical Characteristics Notes: Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue Operation RM5261A only guaranteed with Phase Lock Loop Enabled. JTAG Clock Period tJTAGCKP ilic ModeClock Period tModeCKP RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary Data Output2,3 mode14.13 105,6 (fastest) mode14.13 015,6 (slowest) Parameter1 Symbol Conditions Units 4.75 5.75 Units Boot-Time Interface Parameters Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue Mode Data Hold ilic Mode Data Setup Parameter Symbol Capacitive load maximum output timings Minimum output timings theoretical load condition-untested. Data Output timing applies signal pins whether tristate output only. Setup Hold parameters apply signal pins whether tristate input only. Only mode 14:13 tested guaranteed. Data shown I/O. (250 speeds only) derate times Timings measured from 0.425 VCCIO clock 0.425 VCCIO signal 3.3V I/O. Timings measured from 0.48 VCCIO clock 0.48 VCCIO signal 2.5V I/O. Notes: Data Hold4 tfall above table Data Setup4 tDS6 trise above table SysClock cycles SysClock cycles Speed System Interface Parameters1 RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary tRise tFall ±tJitterIn Figure Input Timing SysClock Figure Output Timing SysClock Data Data tDOmin Data Data Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue ilic Data tDOmax 10.1 System Interface Timing (SysAD, SysCmd, ValidIn*, ValidOut*, etc.) tHigh tLow SysClock Figure Clock Timing Timing Diagrams RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary (19.80) -D1.25 (0.049) 0.20 (0.008) C0.864 (4X) (1.25 (0.049)) 28.10 (1.106) 27.90 (1.098) 0.20 (0.008) 0.05 (0.002) -D0.27 (0.011) 0.17 (0.007) 0.08 (0.003) 0.20 (0.008) 0.05 (0.002) 28.10 (1.106) 27.90 (1.098) VIEW (25.50 (1.004)) 30.85 (1.215) 30.35 (1.195) 4.07 (0.160 MAX. 3.60 (0.142) 3.20 (0.126) BASE PLANE -C0.102 (0.003) SEATING PLANE 5°-16° SIDES MIN. R0.13 (0.005) MIN. R0.13 (0.012) 0.13 (0.005) 0°-7° GAGE PLANE 0.25 (1.30 (0.051)) DETAIL 0.75 (0.029) (208X) 0.50 (0.020) DETAIL AFTER PLATING 0.20 (0.008) 0.09 (0.004) 0.25 (0.010) MIN. Notes Controlling dimensions: millimeters. Dimensions inches shown parentheses. Dimensions tolerancing ANSI Y14.5 1982. Datum plane located mold parting line coincident with lead exits plastic body bottom parting line. Datums "A-B" determined datum plane "H". determined seating plane "C". These dimensions determined datum plane "H". Dimensions include mold protrusion. Allowable protrusion 0.25/0.10" side. Lead width does include damber protrusion. Allowable damber protrusion shall 0.08 mm/0.003" total excess this dimension maximum material condition. Dambar cannot located lower radius foot. Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue numbers start with continue counter-clockwise when viewed from top. Package dimensions conform JEDEC MS-029(FA-1). ilic 0.50 (0.020) (204X) DATUM PLANE -HDATUM PLANE 30.85 (1.215) 30.35 (1.195) (25.50 (1.004)) 0.20 (0.008) Packaging Information RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary VCCIO SysAD4 SysAD36 SysAD5 SysAD37 VCCInt SysAD6 SysAD38 VCCIO SysAD7 SysAD39 SysAD8 SysAD40 VCCInt ModeClock JTDO JTDI JTCK JTMS VCCIO VCCIO VCCInt VCCInt VCCIO Int0* SysAD52 VCCIO SysCmdP SysAD20 VCCIO SysAD53 SysAD21 SysAD22 SysAD54 VCCInt SysAD23 SysAD55 SysAD24 SysAD56 VCCIO SysAD25 SysAD57 VCCInt SysAD26 SysAD58 SysAD27 SysAD59 VCCIO Int2* Int1* Int3* Int4* Int5* VCCIO VCCIO NMI* ExtRqst* Reset* ColdReset* VCCOK BigEndian ModeIn ilic RdRdy* WrRdy* ValidIn* ValidOut* Release* VCCP VSSP SysClock VCCInt VCCIO SysAD9 SysAD41 VCCIO SysAD10 SysAD42 Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue VCCIO SysAD47 SysCmd8 Function Function Function Function Function SysAD30 SysAD62 VCCIO SysAD31 SysAD63 SysADC2 SysADC6 VCCInt SysADC3 SysADC7 VCCIO SysADC0 SysADC4 VCCInt SysADC1 SysADC5 SysAD0 SysAD32 VCCIO SysAD1 SysAD33 VCCInt RM5261A 208-QFP Package Numerical Pinout RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary SysAD11 SysAD43 VCCInt SysAD12 SysAD44 VCCIO SysAD13 SysAD45 SysAD14 SysAD46 VCCInt SysAD15 SysCmd0 SysCmd1 SysCmd2 SysCmd3 VCCIO SysCmd4 SysCmd5 VCCIO SysCmd6 SysCmd7 SysAD16 SysAD48 VCCInt SysAD17 SysAD49 SysAD18 SysAD50 VCCIO SysAD28 SysAD60 SysAD29 SysAD61 VCCInt VCCIO SysAD19 SysAD51 VCCInt Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue ilic VCCInt VCCIO SysAD2 SysAD34 SysAD3 SysAD35 VCCIO VCCIO Function Function Function Function Function RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary ColdReset* ExtRqst* Int0* Int1* Int2* Int3* Int4* Int5* JTCK JTDI JTDO JTMS ModeClock ModeIn SysAD5 SysAD6 SysAD7 SysAD8 SysAD9 SysAD10 SysAD11 SysAD12 SysAD13 SysAD14 SysAD15 SysAD16 SysAD17 SysAD18 SysAD19 SysAD20 SysAD47 SysAD48 SysAD49 SysAD50 SysAD51 SysAD52 SysAD53 SysAD54 SysAD55 SysAD56 SysAD57 VCCInt VCCInt VCCInt VCCInt VCCInt BigEndian SysAD4 SysAD46 VCCInt Function Function Function Function Function RM5261A 208-QFP Package Alphabetical Pinout VCCInt VCCInt VCCInt VCCInt VCCInt VCCInt VCCInt VCCIO VCCIO VCCIO VCCIO VCCIO VCCIO VCCIO VCCIO VCCIO VCCIO VCCIO VCCIO VCCIO VCCIO SysAD58 SysAD59 SysAD60 SysAD61 SysAD62 SysAD63 SysADC0 SysADC1 SysADC2 SysADC3 SysADC4 SysADC5 SysADC6 SysADC7 SysClock ilic SysAD21 SysAD22 SysAD23 SysAD24 SysAD25 SysAD26 SysAD27 SysAD28 SysAD29 SysAD30 Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary NMI* RdRdy* Release* Reset* SysAD0 SysAD1 SysAD2 SysAD3 SysAD31 SysAD32 SysAD33 SysAD34 SysAD35 SysAD36 SysAD37 SysAD38 SysAD39 SysAD40 SysAD41 SysAD42 SysAD43 SysAD44 SysAD45 SysCmd0 SysCmd1 SysCmd2 SysCmd3 SysCmd4 SysCmd5 SysCmd6 SysCmd7 SysCmd8 SysCmdP ValidIn* ValidOut* VCCInt VCCInt VCCIO VCCIO VCCIO VCCIO VCCIO VSSP WrRdy* VCCIO VCCIO Function Function Function Function Function Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue ilic VCCInt VCCIO VCCIO VCCIO VCCIO VCCOK VCCP RM5261AMicroprocessor with 64-Bit System Data Sheet Preliminary RM5261A -123 Temperature Grade: (blank) commercial Industrial Package Type: MQFP with internal heat spreader Device Maximum Speed Device Type 0.18 micron process geometry Valid Combinations RM5261A-250-H RM5261A-400-H Proprietary Confidential PMC-Sierra, Customer's Internal Document PMC-2002240, Issue ilic RM5261A-350-H RM5261A-300-HI (contact sales prior design) RM5261A-300-H Ordering Information Other recent searchesUVP-2088 - UVP-2088 UVP-2088 Datasheet UVP-2X88 - UVP-2X88 UVP-2X88 Datasheet SHD117136 - SHD117136 SHD117136 Datasheet MUR1540 - MUR1540 MUR1540 Datasheet MUR1550 - MUR1550 MUR1550 Datasheet MUR1560 - MUR1560 MUR1560 Datasheet LTC3851 - LTC3851 LTC3851 Datasheet LH5116-10F - LH5116-10F LH5116-10F Datasheet KPDA02-102 - KPDA02-102 KPDA02-102 Datasheet IM14400-1 - IM14400-1 IM14400-1 Datasheet FJT44 - FJT44 FJT44 Datasheet
Privacy Policy | Disclaimer |