Setting up programming environment

Local installation

On your own laptop/workstation and depending on which programming language you would like to do the hands-on labs, you’ll need either Python + mpi4py + numpy (see the mpi4py github repo and mpi4py PyPI repo) OR a C, C++, or Fortran compiler and an MPI library, we recommend the latest version of either open-mpi or mpich.

Using HPC cluster

VSC-5 – during an instructor-led course

  • Before an instructor-led course on vsc5 you will get an “ASC login” mail with your personal login credentials for the course.

  • Login via browser at: https://jupyterhub.vsc.ac.at with your username and password from the email. You will get an OTP (one-time password) as a text message on your mobile phone. If you do not get the OTP the instructor or the teaching assistants can assist you.

  • Hit the red “Start” button on the bottom with keeping the default settings in the JupyterHub.

  • Once your Server is up and running , you’ll see the hands-on labs in the left menu. Double click the “MPI” course folder and then you can choose your programming language “C”, “Fortran”, or “Python” where you can double click on the individual notebooks to open them.

  • If you prefer to work with the code snippets directly, you can open a terminal in the JupyterHub and cd to ~/MPI/tasks/ and choose one of the subfolders C, Fortran, or Python, where you’ll find subfolders for the labs.

  • At the end of each course day please shut down your server properly: In the upper left corner of a Jupyterhub window click on “File” and in the drop-down click on “Hub Control Panel”. On the next page click the “Stop my Server” button and after that “Logout” in the upper right corner of your browser window. Close all open Jupyterhub tabs.

VSC-5 – outside of a course (self-paced learning)

  • You need a username and password valid for vsc5.

  • Option A: Login via browser at: https://jupyterhub.vsc.ac.at

  • Select the profile “VSC-5 (conda python env)” and the Conda env “Conda jupyterhub-vsc-v3 (Python 3.11.8)” with “4 cores”, “15GB”, and select your maximum running time. Then you can press the red “Start” button.

  • Once your Server is up and running, you can open a terminal to copy the course material. In the terminal do: cd ~; cp ~training/MPI MPI

  • You can now either work with the provided Jupyter notebooks or continue in the terminal and cd to ~/MPI/tasks/ and choose one of the subfolders C, Fortran, or Python, where you’ll find subfolders for the labs.

  • You will see the hands-on labs in the left menu in the JupyterHub (maybe you have to navigate to your home directory first). Double click the “MPI” course folder and then you can choose your programming language “C”, “Fortran”, or “Python” where you can double click on the individual notebooks to open them.

  • Option B: Login via ssh

  • Copy the course material: cd ~; cp ~training/MPI MPI

  • Load an MPI module, e.g., module load openmpi/4.1.6-gcc-12.2.0-exh7lqk

  • Cd to ~/MPI/tasks/ and choose one of the subfolders C, Fortran, or Python, where you’ll find subfolders for the labs.

  • The exercises are lightweight and can be done directly on a login node of vsc5.

Other systems

  • You need a username and password to login.

  • Load an MPI module, e.g., module load openmpi (checkout the module names available with module avail)

  • Get the course material, e.g., git clone https://gitlab.tuwien.ac.at/vsc-public/training/mpi

  • Go to ~/MPI/tasks/ and choose one of the subfolders C, Fortran, or Python, where you’ll find subfolders for the labs.