Let f be the fraction of a computation (in terms of time) that is parallelizable, P the number of processors in the system, and \(s_p\) the speed up achievable in comparison with sequential execution – then the \(s_p\) can be calculated using the relation :
UGC NET - 2012,
Answer: C
If a fraction f of the computation is parallelisable and P processors are used, the parallel part takes time f/P, while the serial part, 1 - f, remains. The speed up is the old time divided by the new time
\[ s_p = \frac{1}{(1 - f) + f/P} \]which is Amdahl's law. The first option has the term f/P subtracted, which would make the speed up larger than P, and the second and fourth options do not reduce to Amdahl's formula. The keyed answer is the third option.
Please log in or register to report this question.