4_14_24_3Chapter 4The ProcessorChapter 4 — The Processor — 2Introduction CPU performance factors Instruction count CPI and Cycle time We will examine two MIPS implementations A single-cycle implementation A pipelined version Simple subset, shows most aspects Memory reference: lw, sw Arithmetic/logical: add, sub, and, or, slt Control transfer: beq, j§4.1 IntroductionChapter 4 — The Processor — 3Instruction Execution PC instruction memory, fetch instruction Register numbers register file, read registers Depending on instruction class Use ALU to calculate Arithmetic result Memory address for load/store Branch target address Access data memory for load/store PC target address or PC + 4Chapter 4 — The Processor — 4CPU OverviewChapter 4 — The Processor — 5Multiplexers Can’t just join wires together Use multiplexersChapter 4 — The Processor — 6ControlChapter 4 — The Processor — 7Logic Design Basics§4.2 Logic Design Conventions Information encoded in binary Low voltage = 0, High voltage = 1 One wire per bit Multi-bit data encoded on multi-wire buses Combinational element Operate on data Output is a function of input State (sequential) elements Store informationChapter 4 — The Processor — 8Combinational Elements AND-gate Y = A & BABYI0I1YMuxS Multiplexer Y = S ? I1 : I0ABY+ABYALUF Adder Y = A + B Arithmetic/Logic Unit Y = F(A, B)Chapter 4 — The Processor — 9Sequential Elements Register: stores data in a circuit Uses a clock signal to determine when to update the stored value Edge-triggered: update when Clk changes from 0 to 1DClkQClkDQChapter 4 — The Processor — 10Sequential Elements Register with write control Only updates on clock edge when write control input is 1 Used when stored value is required laterDClkQWriteWriteDQClkChapter 4 — The Processor — 11Clocking Methodology Combinational logic transforms data during clock cycles Between clock edges Input from state elements, output to state element Longest delay determines clock periodChapter 4The ProcessorChapter 4 — The Processor — 2Building a Datapath Datapath Elements that process data and addressesin the CPU Registers, ALUs, mux’s, memories, We will build a MIPS datapath incrementally Refining the overview design§4. Building a DatapathChapter 4 — The Processor — 3Instruction etch2-bit registerIncrement by 4 for next instructionChapter 4 — The Processor — 4ormat Instructions Read two register operands Perform arithmetic/logical operation Write register resultChapter 4 — The Processor — 5LoadStore Instructions Read register operands Calculate address using 1-bit offset Use ALU, but sign-extend offset Load: Read memory and update register Store: Write register value to memoryChapter 4 — The Processor — 6Branch Instructions Read register operands Compare operands Use ALU, subtract and check ero output Calculate target address Sign-extend displacement Shift left 2 places (word displacement) Add to PC + 4 Already calculated by instruction fetchChapter 4 — The Processor — 7Branch Instructionsustre-routes wiresSign-bit wire replicatedChapter 4 — The Processor — 8Composing the Elements First-cut data path does an instruction in one clock cycle Each datapath element can only do one function at a time Hence, we need separate instruction and data memories Use multiplexers where alternate data sources are used for different instructionsChapter 4 — The Processor — 9TypeLoadStore DatapathChapter 4The ProcessorChapter 4 — The Processor — 2ull DatapathChapter 4 — The Processor — 3LU Control ALU used for Load/Store: F = add Branch: F = subtract R-type: F depends on funct field§4.4 A Simple Implementation SchemeALU control Function0000 AND0001 OR0010 add0110 subtract0111 set-on-less-than1100 NORChapter 4 — The Processor — 4LU Control Assume 2-bit ALUOp derived from opcode Combinational logic derives ALU controlopcode ALUOp Operation funct ALU function ALU controllw 00 load word add 0010sw 00 store word add 0010beq 01 branch equal subtract 0110R-type 10 add 100000 add 0010subtract 100010 subtract 0110AND 100100 AND 0000OR 100101 OR 0001set-on-less-than 101010 set-on-less-than 0111Chapter 4 — The Processor — 5The Main Control Unit Control signals derived from instruction0 rs rt rd shamt funct1:2 :02:21 20:1 1:11 10: or 4 rs rt address1:2 2:21 20:1 1:04 rs rt address1:2 2:21 20:1 1:0R-typeLoad/StoreBranchopcode always readread, except for loadwrite for R-type and loadsign-extend and addChapter 4 — The Processor — 6Datapath ith ControlChapter 4 — The Processor — 7Type InstructionChapter 4 — The Processor — 8Load InstructionChapter 4 — The Processor — 9BranchonEqual InstructionChapter 4 — The Processor — 10ControllerR-format Iw sw beqOp0Op1Op2OpOp4OpInputsOutputsRegDstALUSrcMemtoRegRegWriteMemReadMemWriteBranchALUOp1ALUOpOR-format lw sw beqOpcode000000 100011 101011 000100OutputsRegDst 1 0 ALUSrc 0 1 1 0MemtoReg 0 1 RegWrite 1 1 0 0MemRead 0 1 0 0MemWrite 0 0 1 0Branch 0 0 0 1ALUOp1 1 0 0 0ALUOp2 0 0 0 1Chapter 4 — The Processor — 11Implementing umps ump uses word address Update PC with concatenation of op 4 bits of old PC 2-bit jump address 00 Need an extra control signal decoded from opcode2 address1:2 2:0umpChapter 4 — The Processor — 12Datapath ith umps ddedChapter 4 — The Processor — 13 Pipelined Implementation Why isn’t single cycle enough? control is relatively simple CPI is 1, but cycle time must be long enough for every instruction to complete branch instruction versus load instruction loads require instruction fetch, register access, ALU, memory access, register access branches require instruction fetch, register access, ALU and this is for a simplified processor no floating point ops, no multiply or divideUp
View Full Document