(a) Briefly describe the typical 5-stage pipelining in the DLX/RISC architecture. State possible events for load,...

Transcribed Image Text:

(a) Briefly describe the typical 5-stage pipelining in the DLX/RISC architecture. State possible events for load, branch, and register-register ALU instructions in its execution stage (EXE stage). (6 marks) (b) The following code is executed in a DLX pipelined processor. R2, R2, R3 //R2=R2-R3 R1, 20(R2) //load word into R1 //increment R1 by 1 R1, R1, #1 20(R3), R1 R2, R2, #4 //store word from R1 to memory //increment R2 by 4 //R4=R3+R5 R4, R3, R5 SUB LW ADDI SW ADDI ADD • Draw a pipeline timing table to indicate possible pipeline hazards in the execution, and explain why stalling the pipeline is needed to avoid the hazards. • When the machine is equipped with forwarding hardware, re-draw the pipeline timing table indicating where forwardings are needed. Explain how the forwarding techniques are used in this case to achieve a hazard-free operation with the optimal performance. (10 marks) (c) Calculate the speedup gained from the forwarding technique for the code in (b). Show the steps of your calculation and state any assumptions you make. Restrict the result to 2 decimal digits. (4 marks) (a) Briefly describe the typical 5-stage pipelining in the DLX/RISC architecture. State possible events for load, branch, and register-register ALU instructions in its execution stage (EXE stage). (6 marks) (b) The following code is executed in a DLX pipelined processor. R2, R2, R3 //R2=R2-R3 R1, 20(R2) //load word into R1 //increment R1 by 1 R1, R1, #1 20(R3), R1 R2, R2, #4 //store word from R1 to memory //increment R2 by 4 //R4=R3+R5 R4, R3, R5 SUB LW ADDI SW ADDI ADD • Draw a pipeline timing table to indicate possible pipeline hazards in the execution, and explain why stalling the pipeline is needed to avoid the hazards. • When the machine is equipped with forwarding hardware, re-draw the pipeline timing table indicating where forwardings are needed. Explain how the forwarding techniques are used in this case to achieve a hazard-free operation with the optimal performance. (10 marks) (c) Calculate the speedup gained from the forwarding technique for the code in (b). Show the steps of your calculation and state any assumptions you make. Restrict the result to 2 decimal digits. (4 marks)