April 12, 2021 - by CSCS

In this tutorial we will discuss how to use the JupyterHub service on Piz Daint for data analysis, visualization and rapid prototyping of code.

In the first part we will show you how to install Python modules, use virtual environments and install custom kernels into your JupyterLab or notebook environment. We will discuss dask, ipyparallel and mpi4py, and demonstrate NVIDIA's GPU dashboards for basic profiling of GPU code.

The second part will focus on the use of ParaView in Jupyter notebooks, covering topics including migrating from ParaView desktop to ParaView in Jupyter, SMP- and MPI-based parallellism in ParaView within a notebook, and dask, numpy and xarray in ParaView.

The final part will cover the Julia language, providing an overview of the Julia package ecosystem and an introduction to GPU programming with Julia.

This tutorial is targeted at current users of Piz Daint and will be held over Zoom.

For more information please visit the event page >