__author__ = 'ferrard' # --------------------------------------------------------------- # Imports # --------------------------------------------------------------- import time import matplotlib.pyplot as plt import scipy as sp # --------------------------------------------------------------- # Interface - Timing function # --------------------------------------------------------------- def time_them(k, m, *functions): """Times the functions (accepting one argument - n) on k values of n up to m Stops the timing once the function's execution takes: - more then 2 sec - more then 1 sec longer then on previous value of n """ n_values = list(range(1, m)) if m > k: n_values = list(range(1, m, m//k)) results = [] for i in range(len(functions)): print("Testing function " + functions[i].__name__) results_for_f = [] for n in n_values: print("\tInput size " + str(n)) before = time.time() functions[i](n) after = time.time() results_for_f.append(after - before) if results_for_f[-1] > 2 or (len(results_for_f) > 1 and results_for_f[-1] - results_for_f[-2] > 1): break results.append(results_for_f) for i in range(len(functions)): plt.plot(n_values[:len(results[i])], results[i], label=functions[i].__name__) plt.legend() plt.show() # --------------------------------------------------------------- # Interface - try out # --------------------------------------------------------------- def n_sqrt_n(n): res = 0 for i in range(n*int(sp.sqrt(n))): res += 1 return res def n_squared(n): res = 0 for i in range(n*n): res += 1 return res # --------------------------------------------------------------- # Interface - Sum to # --------------------------------------------------------------- def sum_builtin(n): """Sums numbers up to n using built-in function - O(n)""" print(sum(range(n))) def sum_explicit(n): """Sums numbers up to n explicitely - O(n)""" total = 0 for i in range(n): total += i print(total) def sum_analytic(n): """Sums numbers up to n, analytically - O(1)""" print(n*(n + 1)//2) # --------------------------------------------------------------- # Fibonnachi numbers # --------------------------------------------------------------- def fib_n_naive(n): """Naive (recursive) way to compute Fibonacci's numbers. O(F(n))""" if n == 0: return 0 if n == 1: return 1 return fib_n_naive(n - 1) + fib_n_naive(n - 2) def fib_n_efficient(n): """Efficient way to compute Fibonacci's numbers. Complexity = O(n)""" a = 0 b = 1 for i in range(n - 1): c = a + b a = b b = c print(b) return b def fib_n_closed(n): """Closed-form computation Fibonacci's numbers. Complexity = O(n) WRONG! Problems with precision! """ fi = (1 + sp.sqrt(5))/2 res = int(round((fi**n - (-fi)**(-n))/sp.sqrt(5))) print(res) return res # --------------------------------------------------------------- # Sorting # --------------------------------------------------------------- BOUND = 1000 # BOUND = 1000000 # try this bound - the linear sort will much more slow down def sort_selection(n): """Sort n random numbers - using inefficient quadratic sort - O(n^2)""" l = list(sp.random.random_integers(0, BOUND, n)) for i in range(n): for j in range(i + 1, n): if l[i] > l[j]: tmp = l[i] l[i] = l[j] l[j] = tmp def sort_inbuilt(n): """Sorts n random numbers - using efficient inbuilt function - O(n log n)""" l = list(sp.random.random_integers(0, BOUND, n)) l.sort() def sort_linear(n): """Sorts n random numbers bounded in a small range - using efficient linear sort, called Counting sort - O(n)""" l = list(sp.random.random_integers(0, BOUND, n)) counts = [0]*(BOUND + 1) for i in l: counts[i] += 1 counter = 0 for i in range(len(counts)): for j in range(counts[i]): l[counter] = i counter += 1 # --------------------------------------------------------------- # Main # --------------------------------------------------------------- def main(): # time_them(20, 1000, n_sqrt_n, n_squared) # time_them(20, 1000000, sum_builtin, sum_explicit, sum_analytic) # time_them(20, 100, fib_n_naive, fib_n_closed, fib_n_efficient) # time_them(20, 1000, fib_n_closed, fib_n_efficient) # time_them(20, 10000, sort_inbuilt, sort_linear, sort_selection) time_them(20, 1000000, sort_inbuilt, sort_linear) if __name__ == '__main__': main()