The following program is a simple example of performing parallel computing using IPython. It calculates the power on the cluster in a single engine or in parallel in all the engines. Before executing this program, you are advised to check whether the zmq
package is installed or not, as it is required.
To run these programs in IPython, first start the IPython cluster using the ipcluster start --n=4 --profile=testprofile
command. It will create the ipcontroller-engine.json
and ipcontroller-client.json
files in the <userhome>/.ipython/profile_testprofile/security
directory. These files will be searched through when we create the client by passing profile='testprofile'
. If we create the client using parallel.Client()
, then it will search for JSON files in the profile_default
folder.
First, the program defines a function to calculate power, and then it creates a client using a test profile. To call a Python function in an engine, we can use the apply...