Note

This documentation is for a development version of IPython. There may be significant differences from the latest stable release (1.2.1).

Using IPython for parallel computing

• Overview and getting started
  • Examples
  • Introduction
  • Architecture overview
  • Getting Started
• Starting the IPython controller and engines
  • General considerations
  • Using ipcluster
  • Configuring an IPython cluster
  • IPython on EC2 with StarCluster
  • Using the ipcontroller and ipengine commands
• IPython’s Direct interface
  • Starting the IPython controller and engines
  • Creating a DirectView instance
  • Quick and easy parallelism
  • Calling Python functions
  • Moving Python objects around
  • Other things to look at
• Parallel Magic Commands
  • The Magics
  • Multiple Active Views
  • Engines as Kernels
• The IPython task interface
  • Starting the IPython controller and engines
  • Creating a LoadBalancedView instance
  • Quick and easy parallelism
  • Dependencies
  • Retries and Resubmit
  • Schedulers
  • More details
• The AsyncResult object
  • Beyond multiprocessing’s AsyncResult
  • Metadata
  • Map results are iterable!
• Using MPI with IPython
  • Additional installation requirements
  • Starting the engines with MPI enabled
  • Actually using MPI
• IPython’s Task Database
  • Enabling a DB Backend
  • Using the Task Database
  • Example Queries
  • Cost
• Security details of IPython
  • Process and network topology
  • Application logic
  • Secure network connections
  • Specific security vulnerabilities
  • Other security measures
  • Summary
• Getting started with Windows HPC Server 2008
  • Introduction
  • Setting up your Windows cluster
  • Installation of IPython and its dependencies
  • Starting an IPython cluster
  • Performing a simple interactive parallel computation
• Parallel examples
  • 150 million digits of pi
  • Conclusion
• DAG Dependencies
  • Why are DAGs good for task dependencies?
  • A Sample DAG
• Details of Parallel Computing with IPython
  • Caveats
  • Running Code
  • Views
  • Data Movement
  • Results
  • Querying the Hub
  • Controlling the Engines
  • Synchronization
  • Map
  • Decorators and RemoteFunctions
  • Dependencies
• Transitioning from IPython.kernel to IPython.parallel
  • Processes
  • Creating a Client
  • Apply
  • MultiEngine to DirectView
  • Task to LoadBalancedView