SUPR
AI for electron diffraction
Dnr:

NAISS 2024/23-548

Type:

NAISS Small Storage

Principal Investigator:

Magnus Röding

Affiliation:

Chalmers tekniska högskola

Start Date:

2024-09-23

End Date:

2025-04-01

Primary Classification:

10105: Computational Mathematics

Webpage:

Allocation

Abstract

This proposal is made in connection to the project NAISS 2024/5-125 and is related to ML/AI aspects covered by the part of that project that concerns position averaged convergent beam electron diffraction (PACBED), a relatively new technique for characterizing materials, in particular 2D materials. Specifically, it concerns storage needs for developing methods for using PACBED for predicting stress and strain (compression/expansion in different directions due to mechanical load), sample tilts and rotation. For that part of NAISS 2024/5-125, storage needs have increased substantially compared to the expected needs, because due to the high amounts of data, transferring everything between local storage and C3SE storage is infeasible, leading to both raw simulation data and training data generated from those simulations have to be stored on C3SE storage simultaneously. The 500 GiB currently available NAISS 2024/5-125 is hence a bottleneck for studying more cases concurrently. and will remain so for the duration of the project. Therefore, we apply for an *additional* 1500 GiB to the current my/our "/mimer/NOBACKUP/groups/naiss2023-5-44" storage connected to NAISS 2024/5-125. All other aspects of the project NAISS 2024/5-125 remain the same.