Question: Suppose we have a 1X computer that can execute n statements in time t. Assume that an algorithm which is O(f(n)) requires the execution of
Suppose we have a 1X computer that can execute n statements in time t. Assume that an algorithm which is O(f(n)) requires the execution of a number of statements proportional to f(n).
For example, if A1 ε (n2), then assume that for a problem of size n, the algorithm will execute c*n2 statements (where c is a constant). For simplicity, you can assume c = 1.
a. If we run algorithm A2 ε O(n2) with problem size = n on the 1X computer, how long will it take to run in terms of t?
b. If we run algorithm A3 ε O(nlgn) with problem size = n on the 1X computer, how long will it take to run in terms of t?
c. Suppose an algorithm A4 runs on the 1X computer in time t. When we add 1 more element of data, the same algorithm runs in time 2t. Find the complexity (Big O) of A4.
d. If we want A2 to finish in time t with problem size = n, determine how many times faster then our current 1X computer we need to purchase.
e. How long will it take our 1X computer to run 1,000,000 statements?
Step by Step Solution
There are 3 Steps involved in it
The detailed answer for the above question is provided below a If we run algorithm A2 On2 with probl... View full answer
Get step-by-step solutions from verified subject matter experts
