Computer experiment in condensed matter physics
Abstract
In 2019, the major themes of our
computations will be a) melting under extreme conditions of pressure and temperature;
b) partitioning of light elements between liquid and solid iron at extreme conditions
c) creating training sets for metals and development of machine-learned (ML) potentials;
d) application of ML potentials for studying metals (Fe, Mo, Ti).
In all these topics the most time consuming part is molecular dynamics simulations,
especially the ab initio molecular dynamics. All the topics, even when seem unrelated, are critical for the study of mineral physics of the Earth core, the topic funded by the VR. The group of users consist of four persons, however the number of collaborators that will use SNIC resources will be twice as high at different stages of the project. The requested time (800 000 hours/month) is at the lower side what we actually need. We applied for this amount taking into account the needs of the Swedish HPC community.