Numpy: speed of np.ptp() vs np.max()-np.min()

gistfile1.txt

In [1]: import numpy as np

In [2]: a = np.random.random(1e2)

In [3]: timeit a.ptp()
100000 loops, best of 3: 8.3 us per loop

In [4]: timeit a.max()-a.min()
100000 loops, best of 3: 9.31 us per loop


In [5]: a = np.random.random(1e5)

In [6]: timeit a.ptp()
10000 loops, best of 3: 132 us per loop

In [7]: timeit a.max()-a.min()
10000 loops, best of 3: 134 us per loop


In [8]: a = np.random.random(1e8)

In [9]: timeit a.ptp()
10 loops, best of 3: 162 ms per loop

In [10]: timeit a.max()-a.min()
10 loops, best of 3: 161 ms per loop