Parallel Python with IPCluster

Since we use Python software in our class, I would like to talk about Parallel Python with IPCluster ..

IPCluster provides an easy way to do parallel computing in Python. It allow to connect and control Python processes running on a multiple compute cores, either locally or across a cluster. 

Component of IPcluster..

1. Controller: the central process that manages all communications and dispatching of messages and data
2. Client: the user process that connects to the Controller and issues commands or requests results
3. Engine: a "slave" process that responds to commands sent by the Client (via the Controller)

Last but not least, a typical cluster will consist at least one of each of these components, however it is possible to have a "detached" cluster with no clients that is either idle or processing a task queue, and clients connect temporarily to monitor the cluster state, submit tasks (commands), or retrieve results, before disconnecting, and leaving the cluster to persist.

Thanks for reading :)

source :http://continuum.io/blog/ipcluster