- Contact us
Jul 06, 2021 to Jul 09, 2021
High-throughput computing (HTC) is emerging as an effective methodology in computational materials science for the discovery of novel materials, or of new properties of existing materials [Franceschetti:1999,Johannesson:2002,Curtarolo:2003,Curtarolo:2013]. Its adoption is spreading rapidly at the point that HTC is becoming essential to many diverse applications. Just to name a few examples, HTC has been employed to screen materials and optimize properties such as piezoelectricity [Armiento:2011], thermoelectricity and thermal transport [Toher:2014,Bhattacharya:2015], to study new materials for batteries [Hautier:2011,Liu:2015] or water splitting [Castelli:2014], for the prediction of crystal structure [Pickard:2011], or to reconstruct the free-energy surface [Stecher:2014]. Even if efficient simulation tools and enough computer power were available, however, running a high-throughput investigation is not trivial. A typical study may require to run and analyze tens of thousands of calculations, making it essential to rely on automated tools for the submission and management of calculations.
This tutorial will introduce young researchers to HTC, with hands-on tutorials based on the open-source high-throughput platform AiiDA (http://www.aiida.net, [Pizzi:2016]).
The AiiDA framework for workflow management and provenance tracking is backed by a significant community of users and developers, and has interfaces to more than 20 materials science codes (see [aiida-registry]), including to the ab initio codes Quantum ESPRESSO, VASP, cp2k, Castep, Siesta, Fleur, Crystal, NWChem, Wannier90, and Yambo. AiiDA's permissive open source license (MIT) enables participants to use it both in academic and commercial settings. By virtue of its general design and flexible plugin system, AiiDA is easily extended to new codes and new use cases.
This 3-day tutorial will focus on using AiiDA as a tool for HTC: running simulations, retrieving and querying results, using them for obtaining relationships between computed quantities. Moreover, participants will be taught how to install AiiDA on their laptops or workstations, so that at the end of the tutorial they will be ready to deploy AiiDA and use it for their scientific workflows. Discussion sessions, as well as a poster session and an open-mic session are scheduled to foster interaction between the participants, discuss and answer possible questions, and highlight important aspects of the methodology of HTC.
[Armiento:2011] R. Armiento et al., Phys. Rev. B 84, 014103 (2011).
[Bhattacharya:2015] S. Bhattacharya, G.K.H. Madsen, Phys. Rev. B 92, 085205 (2015).
[Castelli:2014] I.E. Castelli et al., Top. Catal. 57, 265 (2014).
[Curtarolo:2003] S. Curtarolo et al., Phys. Rev. Lett. 91, 135503 (2003).
[Curtarolo:2013] S. Curtarolo et al., Nat. Mater. 12, 191 (2013).
[Franceschetti:1999] A. Franceschetti, A. Zunger, Nature 402,