- #1
Aviato78
- 3
- 0
- Homework Statement
- Have a question related to the DLX 5 stage pipelined processor equipped with the forwarding technique. Have to draw pipeline execution drawing table of the instruction sequence and calculate the performance gain achieved by pipeling in terms of speed-up. Honestly confused on how to approach this question. Any help would be appreciated.
- Relevant Equations
- Equations I used in regards to finding the speed-up was Amdahls formula. Where speedup = old execution time / new execution time.
A 5-stage DLX pipelined processor is equipped with the
forwarding technique. The following code is executed in the
processor.
XOR R5, R5, R6 //R5=R5 XOR R6 (XOR: logic operation)
LW R1, 20(R5) //load word into R1
ADDI R1, R1, #2 //increment R1 by 2
SW 20(R5), R1 //store the content in R1 to memory
ADDI R5, R5, #1 //increment R5 by 1
OR R6, R6, R5 //R6=R6 OR R5 (OR: logic operation)
Draw the pipeline execution timing table of the instruction
sequence given above, with marked stalls and forwardings needed
to achieve a hazard-free pipeline for the most efficient performance
in terms of average CPI (the number of clock cycles used for
executing one instruction).
Calculate the performance gain achieved by the pipelining in terms
of speed-up. State any assumptions made in your calculation.
forwarding technique. The following code is executed in the
processor.
XOR R5, R5, R6 //R5=R5 XOR R6 (XOR: logic operation)
LW R1, 20(R5) //load word into R1
ADDI R1, R1, #2 //increment R1 by 2
SW 20(R5), R1 //store the content in R1 to memory
ADDI R5, R5, #1 //increment R5 by 1
OR R6, R6, R5 //R6=R6 OR R5 (OR: logic operation)
Draw the pipeline execution timing table of the instruction
sequence given above, with marked stalls and forwardings needed
to achieve a hazard-free pipeline for the most efficient performance
in terms of average CPI (the number of clock cycles used for
executing one instruction).
Calculate the performance gain achieved by the pipelining in terms
of speed-up. State any assumptions made in your calculation.