Performing parallel computing with IPython
IPython provides a highly declarative framework for parallel computing. Here, we will take an introductory look at it.
Getting ready
This will require IPython. You have to download and prepare the data, as shown in the first recipe. This recipe will not work in the provided Docker container. It's recommended that you have at least a broader overview of the IPython parallel architecture at http://ipython.org/ipython-doc/dev/parallel/parallel_intro.html#architecture-overview.
You will need to start the IPython parallel framework. For this, while inside the directory where you downloaded the data, which is also where you will have to run the recipe code, do in the shell:
ipcluster start -n 4
This will start the controller with four local engines. Make sure that the Python environment running the cluster is the same as the Python environment, where you will run the recipe.
As usual, this is available in the 08_Advanced/IPythonParallel.ipynb notebook.
