691 views
1 votes
1 votes
consider unpipelined machine with 12 ns clock cycles. It uses 4 cycles for alu operation and branches, whereas 5 cycles for memory operations. Assume that the relative frequencies of these operations are 30% 20% 20%.Pipelining overhead is 1 ns. The speedup of the pipeline is:
2.0  3.1  2.8  3.5

1 Answer

Best answer
3 votes
3 votes

CPI = $\frac{.3*4+.2*4+.2*5}{.3+.2+.2}$ = 4.2 
clock cycle timenon-pipeline = CPI*Clock cycle = 4.2*12 = 50.4

CPIpipeline = 1 
Clock cycle time = 1*(12+1)  = 13 

S = $\frac{50.4}{13}$ = 3.87 

selected by

Related questions

1 votes
1 votes
1 answer
1
Alina asked Dec 21, 2018
24,314 views
Non pipelined system takes 130ns to process an instruction . A program of 1000 instructions is executed in non pipelined system. Then same program is processed with proce...
1 votes
1 votes
2 answers
2
Alina asked Jan 1, 2019
2,976 views
Pipeline system has 4 stages and each stage takes 10ns.30% instructions are branch instructions .each branch instruction introduces delay of 3 cycle. What is speed up fac...
0 votes
0 votes
0 answers
3
Alina asked Dec 21, 2018
888 views
Non pipelined system takes 130 ns to process an instruction. A program of 1000 instructions is executed in non pipelined system. Then same program is processed with proce...
0 votes
0 votes
0 answers
4
eyeamgj asked Dec 7, 2018
243 views
https://gateoverflow.in/18090/isro2011-41IN THIS QUESTION NOT GETTING ARJUN SIR’S ANSWER ….WHY 16 IS WRONG BY DOING LIKE SPEEDUP =OLD/NEW