Research Software Engineering in Micromagnetics

Abstract

Software is fundamental to research. Seven out of ten researchers in the United Kingdom report that their work would be impossible without it [1]. Yet the scientific community faces a credibility crisis, and this is publicly known. The crisis is a multistakeholder problem where no single solution will suffice. It has been shown that code quality is strongly related to the quality of the scientific results.
The work presented in this thesis pursues a two-fold strategy. On one hand, in response to the problem just outlined, it explores concepts from software engineering impacting quality in an academic context. This includes version control, software documentation, testing, continuous integration, or software distribution. To render this exploration concrete, a field of study was chosen and two case studies were performed. Like most of the computational sciences, micromagnetics benefits greatly from fast and accurate numerical tools. Thus on the other hand, this work contributes to the body of methodology for micromagnetic simulations.
Two micromagnetic simulators have been developed in preparation of this thesis. One simulator is a finite element code called FinMag. The author created the initial implementation. The other simulator is a finite difference code called Fidimag. For Fidimag, the author participated in the initial implementation. The continuing development of these two packages is now a collaborative effort with many developers.
The two simulators enabled published, novel, reproducible research on magnonics and the study of magnetic skyrmions. Further, the outcome of two numerical studies increasing the performance of the micromagnetic simulators are presented. [1] ”S.J. Hettrick et al, UK Research Software Survey 2014”, DOI:10.5281/zenodo.1183562

More information

Identifiers

ORCID for Hans Fangohr: orcid.org/0000-0001-5494-7193

Catalogue record

Export record