We are pleased to announce the third ATLAS GPU training series, which will be hosted by NVIDIA.
It will be run over the course of 3 days, with the first two covering the basics of GPU architecture and programming methodologies using CUDA. These will take place in an interactive format using the NVIDIA training cluster. The last day will be devoted to programming and running on GPUs from Python.
This interactive tutorial will be remote only, broadcast over Zoom. A recording will be made of the sessions.
Registration is only necessary if you need access to the NVIDIA training cluster for the interactive exercises. If you have access to a machine with an NVIDIA GPU, you can run the exercises yourself. This can either be done by installing the software to run a jupyter notebook, or just executing the commands by hand.
You can test your jupyter notebook infrastructure by trying out the exercises from last year available here: https://indico.cern.ch/event/1154647/attachments/2445872/4192611/JupyterNotebook.tgz
