704 views

1 Answer

Best answer
2 votes
2 votes

Use Amdahl's Law which states :

Effective speedup  =   1 / (1 - f + f / Speedup)

where (1-f) : Fraction of sequential execution

          f  : Fraction of execution with enhancement(Here enhancement is use of multiple CPUs)

        Speedup : Speedup used in parallel part

So according to the question :

          80  =  1  /  (1 - f + f / 100)

 ==>   f    =   395 / 396

 ==>   f    =   0.9975

 ==>   (1 - f) [Sequential part]    =  1 - 0.9975

 ==>   (1 - f) [Sequential part]    =   0.0025

 ==>  Sequential part    =  0.25 %

selected by

Related questions

3 votes
3 votes
1 answer
1
0 votes
0 votes
2 answers
2
farmanahmed888 asked Jan 6, 2022
553 views
Which approach is correct?ASSIGNMENT LINK : https://nptel.ac.in/content/storage2/courses/106105163/ASSIGNMENT-SOLUTION-WEEK11.pdfnptel assignment q-1 Caption
3 votes
3 votes
0 answers
3
thepeeyoosh asked Jan 13, 2018
717 views
Actually, In this problem what will, we consider getting the answer(upper bound or lower bound) and why?
7 votes
7 votes
1 answer
4
akb1115 asked Oct 12, 2017
1,429 views
Consider a non-pipelined processor design which has a cycle time of 15ns and average CPI of 1.6. The maximum speedup pipelined processor can get by pipelining it into 5 s...