Answers are grounded in this paper and cite their sources.
A hybrid finite element method/boundary element method (FEM/BEM) is a standard approach for calculating the magnetostatic potential within micromagnetics [D. Fredkin and T. Koehler, IEEE Trans. Magn. 26, 415 (1990)]. This involves dealing with a dense N×N matrix Bij, with N being the number of mesh surface nodes. In order to apply the method to ferromagnetic structures with a large surface, one needs to apply matrix compression techniques on Bij. An efficient approach is to approximate Bij by hierarchical matrices (or H matrices). We have used HLIB [http://www.hlib.org], a library containing implementations of the hierarchical matrix methodology, together with the micromagnetic finite element solver NMAG in order to optimize the hybrid FEM/BEM. In this article we present a study of the efficiency of algorithms implemented in HLIB concerning the storage requirements and the matrix assembly time in micromagnetic simulations.
No takes yet. Share an insight, caveat, or question.
Knittel et al. (2009) studied this question.
Synapse has enriched 5 related papers. Consider them for comparative context: