Speaker
Description
This presentation aims at describing the experience of a Lattice QCD medium-size research group to develop a research program involving access to large scale HPC systems in Europe: from local clusters, to national resources, and ending with EuroHPC scale.
The presentation intends to be generic enough from the application point of view, but Lattice QCD will be used as a guiding example. We will describe the level of resources required to develop a competitive simulation project from the proof of concept, development, benchmarking and scalability tests points of view. In doing so we will describe the main technical and policy challenges encountered, and discuss possible paths and ideas for improvement.
We will also describe the additional challenges posed by the fact that the research team is distributed across several countries in Europe. Those challenges go well beyond accessing CPU resources for simulation, spanning also to the data management such as distributed access and long term preservation.
