1. CDA 3101 Fall 2013 Introduction to Computer Organization Multicycle Datapath 9 October 2013

2. Review Construction of the Datapath –Determine instruction types R-format Conditional Branch Unconditional Branch Load/store –Develop datapath modules (RF, ALU, Memory, SignExt) –Connect modules to form composite datapath Single-Cycle Datapath –All instructions take one cycle (CPI = 1) –Cycle time dictated by circuit settling time –All operations take time of slowest operation (load) 

3. Overview – Multicycle Datapath Each instruction has multiple stages Each stage takes one cycle 1.Instruction fetch 2.Instruction decode / Data fetch 3.ALU ops / R-format execution 4.R-format completion 5.Memory access completion Each stage can re-use hardware from previous stage More efficient use of hardware and time >> New Hardware required to buffer stage output >> New Muxes required for hardware re-use >> Expanded Control for new hardware All instructions use these

4. Recall: Simple Datapath PC Instruction memory +4 rt rs rd Registers ALU Data memory imm Data Address Controller Opcode, funct Address Instruction ° Datapath is based on register transfers required to execute instructions ° Control causes the right transfers to happen

5. Recall: R-format Datapath Format: opcode r3, r1, r2 Result Zero ALU Read Data 1 Read Data 2 Read Reg 1 Read Reg 2 Write Register Write Data Register Write ALU op Register File 3 Instruction

6. Recall: Load/Store Datapath FetchDecodeExecute

7. Recall: Branch Datapath FetchDecodeExecute

8. Hi-Level View: Multicycle DP Instr. Fetch Instr. Decode/Data Fetch Execute

9. Hi-Level View: Multicycle DP How do we make multicycle datapath (DP)??? 1.Replace 3 ALUs from the single-cycle DP with one ALU 2.Add one multiplexer to select ALU input 3.Add one control line for the ALU input multiplexer New inputs:Constant = 4 [PC + 4] Sign-ext., shifted offset [BTA calc.] 4.Add temporary (buffer) registers: (storage betw.cycles) MDR: Memory Data Register IR: Instruction Register A,B:ALU operand registers ALUout:ALU output register

10. Multicycle DP: The Full Monty

11. Multicycle DP: 1-bit Ctl. Signals

12. Multicycle DP: 2-bit Ctl. Signals

13. Making Sense of Multicycle DP Step 1: Decompose the MC/DP execution sequence into cycles Step 2: Examine which cycles apply to which instructions One-Cycle Steps R-fmt lw sw beq j 1.Instruction Fetch 2.Instruction Decode / Data Fetch 3.ALU ops / R-format Execution 4.R-format Completion 5.Memory Access Completion

14. Multicycle DP: R-format Step 1: Fetch instr. // Store in IR // Compute PC + 4 Step 2: Decode instruction: opcode, rd, rs, rt, funct fields Data fetch: Apply rs, rt to Register File Data Read into A,B buffer registers (ALUin) Step 3: ALU operation ( ALUsrcA, ALUsrcB, ALUop ) ALU output goes into ALUout register Step 4: ALUout register contents written to Register File write input Register number in rd written (Assert: RegWrite,RegDst ) CPI for R-format = 4 cycles

15. Multicycle DP: Store Word ( sw ) Step 1: Fetch instr. // Store in IR // Compute PC + 4 Step 2: Decode instruction: opcode, rs, rt, offset fields Data fetch: Apply rt to Register File => Base address Data Read into A buffer register (Base) SignExt,Shift offset field into B buffer register Step 3: ALU operation ( ALUsrcB, ALUop ) => Base + Offset ALU output goes into ALUout register Step 4: ALUout register contents applied as Memory Address Assert: MemWrite [ ALUout => RegFile] CPI for Store = 4 cycles

16. Multicycle DP: Load Word ( lw ) Step 1: Fetch instr. // Store in IR // Compute PC + 4 Step 2: Decode instruction: opcode, rd, rt, offset fields Data fetch: Apply rt to Register File => Base address Data Read into A buffer register (Base) SignExt,Shift offset field into B buffer register Step 3: ALU operation ( ALUsrcB, ALUop ) => Base + Offset ALU output goes into ALUout register Step 4: ALUout register contents applied as Memory Address Assert: MemRead Step 5: Memory Data Out routed to Register File write input Register number from rd written to (Assert: CPI for Load = 5 cycles

17. Multicycle DP: Cond. Branch Step 1: Fetch instr. // Store in IR // Compute PC + 4 Step 2: Decode instruction: opcode, rs, rt, offset fields Data fetch: Apply rs, rt to Register File BTA calc:SignExt,Shift offset field into B buffer register ALU compose PC, offset => BTA Step 3: ALU operation ( ALUsrcA, ALUsrcB, ALUop ) = compare ALU output present at Zero register causes Control to select BTA or PC+4 CPI for Conditional Branch = 3 cycles

18. Multicycle DP: Jump Step 1: Fetch instr. // Store in IR // Compute PC + 4 Step 2: Decode instruction: opcode, address fields JTA calc:SignExt,Shift offset field [Bits 27-0] Concatenate with PC [Bits 31-28] => JTA Step 3: PC replaced by the Jump Target Address (JTA) PCsource = 10, PCWrite asserted CPI for Jump = 3 cycles

19. Conclusions MIPS ISA: Three instruction formats (R,I,J) One cycle per stage, Different stages per format One-Cycle Steps R-fmt lw sw beq j 1.Instruction Fetch 2.Instruction Decode / Data Fetch 3.ALU ops / R-format Execution 4.R-format Completion 5.Memory Access Completion Challenge: More involved control design

20. Almost Here: Pipelining