Profiling with IPython

In IPython, we can profile small snippets of code using timeit. We can also profile a larger script. We will show both approaches.

How to do it...

First, we will time a small snippet.

1. Timing a snippet.

   Start IPython in pylab mode:

   ipython -pylab

   Create an array containing 1000 integer values between 0 and 1000:

   In [1]: a = arange(1000)

   Measure the time taken for searching "the answer to everything"—42, in the array. Yes, the answer to everything is 42. If you don't believe me please read http://en.wikipedia.org/wiki/42_%28number%29.

   In [2]: %timeit searchsorted(a, 42)
   100000 loops, best of 3: 7.58 us per loop

2. Profile a script.

   We will profile this small script that inverts a matrix of varying size containing random values. The .I attribute (that's uppercase I) of a NumPy array represents the inverse of a matrix:

   import numpy
   
   def invert(n):
     a = numpy.matrix(numpy.random.rand(n, n))
     return a.I
   
   sizes = 2 ** numpy.arange(0, 12)
   
   for n in sizes:
     invert(n)

   We can time this as...