1. 1 Designing a Pipelined Processor In this Chapter, we will study 1. Pipelined datapath 2. Pipelined control 3. Data Hazards 4. Forwarding 5. Branch Hazards 6. Performance

2. 2 Pipelining is Natural Car Wash Station (non-pipelined) Vacuum WashDry Brian Steve Brian Bob Steve Car Wash Station (pipelined) Vacuum Wash Dry Brian Steve Brian Bob Steve Brian Mike Bob SteveBrian Gorge Mike BobSteve Vacuum, Wash, Dry each takes 5 minutes Nonpipelined Car Wash Station serving one customer per 15 minutes Pipelined Car Wash Station after the first customer (it takes 15 minutes), then every 5 minutes finishing service for a new customer

3. 3 What we learn from car wash station Pipelining does not help latency of a single task, but it helps throughput of entire workload Pipeline rate is limited by the slowest pipeline stage Multiple tasks operating simultaneously using different resources Potential speedup = Number of stages Unbalanced lengths of pipe stages reduces speedup Time to fill the pipe and time to drain the pipe -> reduce speedup Stall the pipe for dependences

4. 4 Performance Suppose we execute 100 instruction single cycle implementation –50 ns/cycle x 1 CPI x 100 inst. = 5000 ns multiple cycle implementation –10 ns/cycle x 4.6 CPI (instr. mix) x 100 inst = 4600 ns ideal pipelined processor –10 ns/cycle x (1 CPI x 100 inst + 4 cycle to fill up the pipe) = 1040 ns –here we assume that 5 stages in the pipeline –5 stages : IF, ID, EX, MEM, WB

5. 5 Why pipeline ? Because the resources are there ! Why waste ! Start fetching and executing the next instruction before the current one has completed Remember the performance equation: CPU time = CPI * CC * IC Fetch (and execute) more than one instruction at a time Superscalar processing

6. 6 A Pipelined MIPS Processor Start the next instruction before the current one has completed –improves throughput - total amount of work done in a given time –instruction latency (execution time, delay time, response time - time from the start of an instruction to its completion) is not reduced Cycle 1Cycle 2Cycle 3Cycle 4Cycle 5 IFetchDecExecMemWB lw Cycle 7Cycle 6Cycle 8 sw IFetchDecExecMemWB R-type IFetchDecExecMemWB »clock cycle (pipeline stage time) is limited by the slowest stage »for some instructions, some stages are wasted cycles

7. 7 Single Cycle, Multiple Cycle, vs. Pipeline Multiple Cycle Implementation: Clk Cycle 1 IFetchDecExecMemWB Cycle 2Cycle 3Cycle 4Cycle 5Cycle 6Cycle 7Cycle 8Cycle 9Cycle 10 IFetchDecExecMem lwsw IFetch R-type lw IFetchDecExecMemWB Pipeline Implementation: IFetchDecExecMemWB sw IFetchDecExecMemWB R-type Clk Single Cycle Implementation: lwsw Waste Cycle 1Cycle 2

8. 8 MIPS Pipeline Datapath Modifications What do we need to add/modify in our MIPS datapath? –State registers between each pipeline stage to isolate them

9. 9 Pipelining the MIPS ISA What makes it easy –all instructions are the same length (32 bits) »can fetch in the 1 st stage and decode in the 2 nd stage –few instruction formats (three) with symmetry across formats »can begin reading register file in 2 nd stage –memory operations can occur only in loads and stores »can use the execute stage to calculate memory addresses –each MIPS instruction writes at most one result (i.e., changes the machine state) and does so near the end of the pipeline (MEM and WB) What makes it hard –structural hazards: what if we had only one memory? –control hazards: what about branches? –data hazards: what if an instruction’s input operands depend on the output of a previous instruction?

10. 10 Can pipelining cause us trouble ? Pipeline Hazards : –structural hazards : attempt to use the same resource by two different instructions –data hazards : attempt to use data before it is valid. Such as an instruction depends on results of prior instructions still in the pipeline –control hazards : attempt to execute branch before condition is computed by prior instructions Resolve Pipeline Hazards : –simplest solution : wait until hazards are gone –take action to resolve hazards, the earlier the hazards are solved, the higher the performance the machine can achieve. Pipeline Overhead : –need pipeline registers between stages

11. 11 Why Pipeline? For Performance! I n s t r. O r d e r Time (clock cycles) Inst 0 Inst 1 Inst 2 Inst 4 Inst 3 ALU IM Reg DMReg ALU IM Reg DMReg ALU IM Reg DMReg ALU IM Reg DMReg ALU IM Reg DMReg Once the pipeline is full, one instruction is completed every cycle, so CPI = 1 Time to fill the pipeline

12. 12 Structural Hazard Example MemRegALUMemReg instruction 1 Instruction 2 Instruction 3 instruction 4 time Two instructions want to access the MEM at the same time Structural Hazard ! How to resolve this hazard ?

13. 13 I n s t r. O r d e r Time (clock cycles) lw Inst 1 Inst 2 Inst 4 Inst 3 ALU Mem Reg MemReg ALU Mem Reg MemReg ALU Mem Reg MemReg ALU Mem Reg MemReg ALU Mem Reg MemReg A Single Memory Would Be a Structural Hazard Reading data from memory Reading instruction from memory Fix with separate instr and data memories (I$ and D$)

14. 14 How About Register File Access? I n s t r. O r d e r Time (clock cycles) add $1, Inst 1 Inst 2 add $2,$1, ALU IM Reg DMReg ALU IM Reg DMReg ALU IM Reg DMReg ALU IM Reg DMReg

15. 15 How About Register File Access? I n s t r. O r d e r Time (clock cycles) Inst 1 Inst 2 ALU IM Reg DMReg ALU IM Reg DMReg ALU IM Reg DMReg ALU IM Reg DMReg Fix register file access hazard by doing reads in the second half of the cycle and writes in the first half add $1, add $2,$1, clock edge that controls register writing clock edge that controls loading of pipeline state registers

16. 16 Register Usage Can Cause Data Hazards I n s t r. O r d e r add $1, sub $4,$1,$5 and $6,$1,$7 xor $4,$1,$5 or $8,$1,$9 ALU IM Reg DMReg ALU IM Reg DMReg ALU IM Reg DMReg ALU IM Reg DMReg ALU IM Reg DMReg Dependencies backward in time cause hazards Read After write data hazard

17. 17 Register Usage Can Cause Data Hazards ALU IM Reg DMReg ALU IM Reg DMReg ALU IM Reg DMReg ALU IM Reg DMReg ALU IM Reg DMReg Dependencies backward in time cause hazards add $1, sub $4,$1,$5 and $6,$1,$7 xor $4,$1,$5 or $8,$1,$9 Read after write data hazard

18. 18 Loads Can Cause Data Hazards I n s t r. O r d e r lw $1,4($2) sub $4,$1,$5 and $6,$1,$7 xor $4,$1,$5 or $8,$1,$9 ALU IM Reg DMReg ALU IM Reg DMReg ALU IM Reg DMReg ALU IM Reg DMReg ALU IM Reg DMReg Dependencies backward in time cause hazards Load-use data hazard

19. 19 stall One Way to “Fix” a Data Hazard I n s t r. O r d e r add $1, ALU IM Reg DMReg sub $4,$1,$5 and $6,$1,$7 ALU IM Reg DMReg ALU IM Reg DMReg Can fix data hazard by waiting – stall – but impacts CPI

20. 20 Another Way to “Fix” a Data Hazard I n s t r. O r d e r add $1, ALU IM Reg DMReg sub $4,$1,$5 and $6,$1,$7 ALU IM Reg DMReg ALU IM Reg DMReg Fix data hazards by forwarding results as soon as they are available to where they are needed xor $4,$1,$5 or $8,$1,$9 ALU IM Reg DMReg ALU IM Reg DMReg

21. 21 Another Way to “Fix” a Data Hazard ALU IM Reg DMReg ALU IM Reg DMReg ALU IM Reg DMReg Fix data hazards by forwarding results as soon as they are available to where they are needed ALU IM Reg DMReg ALU IM Reg DMReg I n s t r. O r d e r add $1, sub $4,$1,$5 and $6,$1,$7 xor $4,$1,$5 or $8,$1,$9

22. 22 Forwarding with Load-use Data Hazards I n s t r. O r d e r lw $1,4($2) sub $4,$1,$5 and $6,$1,$7 xor $4,$1,$5 or $8,$1,$9 ALU IM Reg DMReg ALU IM Reg DMReg ALU IM Reg DMReg ALU IM Reg DMReg ALU IM Reg DMReg

23. 23 Forwarding with Load-use Data Hazards ALU IM Reg DMReg ALU IM Reg DMReg ALU IM Reg DMReg ALU IM Reg DMReg ALU IM Reg DMReg Will still need one stall cycle even with forwarding I n s t r. O r d e r lw $1,4($2) sub $4,$1,$5 and $6,$1,$7 xor $4,$1,$5 or $8,$1,$9

24. 24 Branch Instructions Cause Control Hazards I n s t r. O r d e r lw Inst 4 Inst 3 beq ALU IM Reg DMReg ALU IM Reg DMReg ALU IM Reg DMReg ALU IM Reg DMReg Dependencies backward in time cause hazards

25. 25 stall One Way to “Fix” a Control Hazard I n s t r. O r d e r beq ALU IM Reg DMReg lw ALU IM Reg DMReg ALU Inst 3 IM Reg DM Fix branch hazard by waiting – stall – but affects CPI

26. 26 Corrected Datapath to Save RegWrite Addr Need to preserve the destination register address in the pipeline state registers

27. 27 Corrected Datapath to Save RegWrite Addr Need to preserve the destination register address in the pipeline state registers

28. 28 MIPS Pipeline Control Path Modifications All control signals can be determined during Decode –and held in the state registers between pipeline stages Control

29. 29 Other Pipeline Structures Are Possible What about the (slow) multiply operation? –Make the clock twice as slow or … –let it take two cycles (since it doesn’t use the DM stage) ALU IM Reg DMReg MUL ALU IM Reg DM1Reg DM2 What if the data memory access is twice as slow as the instruction memory? –make the clock twice as slow or … –let data memory access take two cycles (and keep the same clock rate)

30. 30 Sample Pipeline Alternatives ARM7 StrongARM-1 XScale ALU IM1 IM2 DM1 Reg DM2 IM Reg EX PC update IM access decode reg access ALU op DM access shift/rotate commit result (write back) ALU IM Reg DMReg SHFT PC update BTB access start IM access IM access decode reg 1 access shift/rotate reg 2 access ALU op start DM access exception DM write reg write

31. 31 Summary All modern day processors use pipelining Pipelining doesn’t help latency of single task, it helps throughput of entire workload Potential speedup: a CPI of 1 and fast a CC Pipeline rate limited by slowest pipeline stage –Unbalanced pipe stages makes for inefficiencies –The time to “fill” pipeline and time to “drain” it can impact speedup for deep pipelines and short code runs Must detect and resolve hazards –Stalling negatively affects CPI (makes CPI less than the ideal of 1)