We have been awarded Academy of Finland funding within the framework of the EuroHPC research program, which aims to support the transition to (pre-)exascale high-performance computing (HPC) platforms and quantum computing, among others. Miguel Caro will lead the ExaFF (“Exascale-ready machine learning force fields“) consortium as Principal Investigator and Consortium Coordinator. This project is a collaboration with Andrea Sand’s group at Aalto University and CSC. The Academy of Finland has granted us 1,011,240 EUR for this project, out of which 435,108 EUR correspond to the Caro group.
The objective of the project is to improve the accuracy and speed of Gaussian approximation potentials through the development of the TurboGAP code and the underlying algorithms. One of the priorities is to adapt them to efficient GPU execution, with a focus on exploiting the computing power provided by the CSC-hosted LUMI pre-exascale HPC system. The faster code and new functionality will be used to model new materials for battery applications and semiconductors under heavy radiation environments (e.g., circuit components aboard satellites).
This is great news for our group, since it secures the resources necessary to continue the development of the TurboGAP code in the short term (the funding runs for 3 years, from 2022 until 2024).
See the funding decision on the Academy of Finland’s website.