Generated by GPT-5-mini| PRACE Training Centre | |
|---|---|
| Name | PRACE Training Centre |
| Formation | 2010s |
| Type | Research infrastructure training |
| Location | Europe |
| Parent organization | Partnership for Advanced Computing in Europe |
PRACE Training Centre The PRACE Training Centre is a European research infrastructure initiative focused on high-performance computing (HPC) training and skills development for researchers, engineers, and technical staff across European Commission, European Research Area, Horizon 2020, Horizon Europe and related programs. It provides coordinated courses, workshops, and summer schools tied to regional HPC facilities such as France-Grilles, ARCHER2, FZJ Jülich, and national initiatives like PRACE aisbl partner sites, engaging communities from CERN, ESA, EMBL to universities including University of Cambridge, TU Delft, and ETH Zurich. The centre links capability computing, code optimisation, and parallel programming with applied domains represented by projects funded under European Strategy Forum on Research Infrastructures and consortia like EuroHPC Joint Undertaking.
The PRACE Training Centre functions as a distributed network delivering standardized curricula in areas such as MPI, OpenMP, GPU programming with NVIDIA, performance profiling using tools from Intel Corporation and AMD, and reproducible workflows connecting to repositories like Zenodo and GitHub. It coordinates accreditation, certification, and trainer exchange programs with research infrastructures including SuperMUC-NG and Piz Daint, and aligns with community standards promoted by Compute Canada and XSEDE analogues. Emphasis is placed on transferable competencies used by researchers affiliated with Max Planck Society, CNRS, CSIC, CINECA, BSC, and national supercomputing centres such as SURFsara and SNIC.
Origins trace to collaborations among major European centres including PRACE aisbl founding partners, BSC Barcelona Supercomputing Center, CEA, EPCC, and DFG-funded groups during initiatives aligned with European Grid Infrastructure. Early milestones involved workshops co-organised with Rutherford Appleton Laboratory and summer schools inspired by events at ETH Zurich and University of Edinburgh. Development phases mirrored policy frameworks from Lisbon Strategy and Bologna Process aiming to harmonise technical training across European Union member states, while drawing on pedagogical models from INRIA and Laboratoire national de métrologie et d'essais collaborations. Expansion included integration with EuroHPC projects and partnerships with private sector vendors such as Cray, IBM, and Arm Holdings.
Programs cover foundational topics including parallel algorithms illustrated by case studies from Large Hadron Collider analyses at CERN, climate modelling used by ECMWF, and bioinformatics pipelines developed at European Bioinformatics Institute and EMBL-EBI. Advanced modules teach heterogeneous computing targeting accelerators used at Oak Ridge National Laboratory and Argonne National Laboratory collaborations, containerisation workflows utilising Docker and Singularity examples from PRACE DIH activities, and machine learning workflows referencing projects at DeepMind and OpenAI for HPC-optimised inference. Course offerings include instructor-led workshops, MOOCs co-branded with edX partners, and hackathons run with organisers such as Mozilla and The Carpentries, adapted for participants from institutions like Imperial College London and Politecnico di Milano.
The network brings together national supercomputing centres including CINECA, CSC – IT Center for Science, LRZ, IT4Innovations, and regional centres such as BSC, PRACE aisbl partners, research organisations like Max Planck Society, CNRS, CSIC, and universities including Sorbonne University, Universität Hamburg, University of Manchester, Heidelberg University, KU Leuven, and University of Warsaw. Industry collaborations span NVIDIA, Intel Corporation, AMD, HPE, Atos, and scientific software vendors like ANSYS and Siemens. International linkages include projects with NSF-backed consortia, DOE laboratories, and advisory interactions with GÉANT and EUDAT.
Training sessions use compute systems such as Piz Daint, SuperMUC-NG, ARCHER2, Frontera (supercomputer), and cloud platforms including Amazon Web Services research grants and Microsoft Azure HPC offerings. Teaching materials rely on repositories hosted by GitHub, archival services like Zenodo, and workflow platforms exemplified by Nextflow and Snakemake used in collaborative exercises with ELIXIR. Access to performance tools comes from collaborations with Intel VTune Amplifier, NVIDIA Nsight, Arm Forge, and energy profiling via instruments used at Jülich Research Centre and CINECA labs.
Outreach efforts include conferences and events co-located with ISC High Performance, Supercomputing Conference (SC), EuroHPC Summit Week, and regional workshops in coordination with RDA and CODATA. The centre has supported researcher upskilling across disciplines from computational chemistry at Max Planck Institute for Coal Research to astrophysics at ESO, contributing to projects like Human Brain Project and GAUSS Centre for Supercomputing collaborations. Impact metrics reference trainee numbers, contributed open educational resources adopted by The Carpentries, and case studies published alongside Nature Computational Science and IEEE conference proceedings.
Governance structures mirror transnational models involving stakeholders such as PRACE aisbl board members, national funding agencies including Agence Nationale de la Recherche, DFG, UK Research and Innovation, Ministerio de Ciencia e Innovación, and frameworks under Horizon Europe and EuroHPC Joint Undertaking. Funding streams combine contributions from European Commission grants, in-kind support from supercomputing centres like CINECA and BSC, and industry partnerships with firms such as HPE and Atos, overseen by advisory committees including representatives from European University Association and scientific societies like ACM and IEEE Computer Society.
Category:European research infrastructure