IPython Notebooks for Parallel Computing

This is simply an example notbook illustrating the "cluster" tab of the ipython notebook. You can use this as a starter project for developing simple parallel codes in a notebook.

First start up a cluster using the "Cluster" tab. Then try these cells.



In [1]:

    
from IPython.parallel import Client
rc=Client()
view = rc[:]
print rc.ids









    



[0, 1, 2, 3]

Use the %px magic to import methods on the nodes



In [2]:

    
%px from pylab import arccos, rand
from numpy import array

Next, use the %%px magic to define a function on the nodes



In [3]:

    
%%px

def f(N):
    y=arccos(1.0-2*rand(N))
    return 2.0*y.sum()/N



In [4]:

    
x = view.map(lambda N: f(N), 10*[1000000]).result
x









    Out[4]:





[3.1405796297219211,
 3.1426586695423637,
 3.1398369967668183,
 3.1420250074438467,
 3.1427624932300482,
 3.1410966404557557,
 3.1431271009553994,
 3.141158465309319,
 3.1426331913328771,
 3.1436420212612775]



In [5]:

    
array(x).sum()/len(x)









    Out[5]:





3.1419520216019627



In [ ]: