XSEDE Gateway Symposium: Jupyter (previously IPython) Notebook in HPC

The Jupyter (previously IPython) Notebook is a browser-based programming front-end that integrates code in diverse programming languages, formatted text, latex equations and embedded plots in the same document. It allows scientists to explore their data, share and reproduce their analysis.

In this talk I'll first introduce how the Jupyter Notebook works, how it can be used in HPC for interactive work, within workflow systems and in Scientific Gateways.
Then I'll introduce Jupyterhub, the multi-user Notebook server, how a plugin I developed (RemoteSpawner) can provide an integration with Torque or SLURM to provide easy access to Supercomputing resources. Finally how from the Notebook it is possible to launch large batch jobs with IPython Parallel.

The purpose of the talk is to get feedback and brainstorm about how to best use Jupyter technologies in HPC.

Bio:
Andrea Zonca has a background in Cosmology, during his PhD and PostDoc he worked on analyzing Cosmic Microwave Background data from the Planck Satellite. In order to manage and analyze large datasets, he developed expertise in Supercomputing, in particular parallel computing in Python and C++. At the San Diego Supercomputer Center he works in the Scientific Computing Applications team, he works on a CUDA/C++ molecular dynamics package, on Cosmology data analysis and on making HPC easier to access with Jupyter and Jupyterhub. Andrea is also a certified instructor of Software Carpentry and teaches automation with bash, version control with git and programming with Python to scientists.