by Carey Sargent, EPFL, NCCR MARVEL
How did you first hear about AiiDA?
We became aware of AiiDA through MARVEL, in its initial phase. It appeared to me to be something interesting, but I had no human resources to go beyond "it's interesting" until I managed to convince a PhD student in our group, Aliaksandr Yakutovich, to spend some time understanding whether or not it was possible to get some support for our calculations on graphene nanoribbons (GNRs) in AiiDA. The support that we immediately received from the entire AiiDA team – going through it step by step – convinced me that it was worth continuing to invest human resources in this direction.
How did you get started with AiiDA?
The revolution came when we obtained additional funds from MARVEL on a project aimed at an advanced AiiDA-based infrastructure for GNRs designed in collaboration with experimentalists. This simple idea was then transformed by Ole Schütt into a clear vision and route towards what we now call AiiDA lab. Again, the initial effort would have led to nothing without the support from the current AiiDA and AiiDA lab team. The apps we have now in AiiDA lab, developed by Edward Ditler and Kristjan Eimre, allow us to focus on science rather than on setting up simulations and give experimentalists direct access to the processed results of calculations, and even to the preparation and submission of simple computational tasks.
How is AiiDA used in the lab?
At present, experimental PhDs and postdocs in our lab submit and monitor computational workflows based on Quantum ESPRESSO and CP2K to compute the electronic properties of GNRs and are able to independently monitor the status of diverse simulations—for example nudged-elastic band, geometry optimizations, scanning tunneling microscopy/spectroscopy, and atomic force microscopy—that have been submitted through AiiDA lab by theoretical colleagues. They can independently visualize or download the results automatically processed by the corresponding AiiDA lab apps. Band structures of GNRs published from our laboratory in the last year were computed by experimentalists and we had one case where we discussed the scientific content of a research work where all density functional theory calculations were done by the experimentalists via AiiDA lab.
(A video on use in the lab can be found here.)
Where do you see AiiDA heading?
As our reliance on the AiiDA lab increases, it becomes evident that computational infrastructure, just like any other infrastructure, needs support. While the MARVEL NCCR with its 12-year life span is a great basis, I believe in the long term we will need to find ways to support targeted development on the institutional level. I am therefore personally exploring the possibility to promote and lead an effort to consolidate the research challenges of our laboratory with AiiDA and other data science developments at Empa.
Low-volume newsletters, targeted to the scientific and industrial communities.Subscribe to our newsletter