Knowledge transfer towards the academic sector
The activities of all the group leaders in MARVEL have already a very robust component of knowledge transfer, both with the development or co-development of key open-source or licensed codes such as CP2K, CPMD, QuantumESPRESSO, ABINIT, BigDFT, ALPS, TRIQS, WANNIER90, Piglet, and with a sustained activity of training workshops and conferences.
In addition, and thanks to the collaboration with the Swiss Platform for Advanced Scientific Computing (PASC), key algorithmic, methodological, and computational improvements are taking place, especially within CP2K, CPMD, QuantumESPRESSO, BigDFT, ABINIT, from optimal performance of sparse matrix libraries, to GPU optimization, to development of new multi-scale code-agnostic efforts.
AiiDA — Automated Interactive Infrastructure and Database for Computational Science — is an open source platform to manage, preserve, and disseminate the simulations, data, and workflows of modern-day computational science developed in the framework of MARVEL.
The AiiDA platform will empower the sharing of materials data, structures, workflows to calculate complex properties, and algorithms for machine learning and data mining.
A regular calendar of AiiDA tutorials will be provided. Two were organised at this time: October 31, 2014 at ETHZ (30 participants) and on February 5, 2015 at the Psi-k/CECAM conference on Materials Design and Discovery (40 participants).
The international collaborations, in particular with the research groups participating in the Materials Genome Initiative, to insure data compatibility and synergies. Natural collaboration arise with the Materials' Project, the Open Quantum Materials Database, the SUNCAT Center for interface science and catalysis, as well as the Crystallography Open Database will be enlarged.