We report on the design, verification and performance of MUMAX3, an open-source GPU-accelerated micromagnetic simulation program. This software solves the time- and space dependent magnetization evolution in nano- to micro scale magnets using a finite-difference discretization. Its high performance and low memory requirements allow for large-scale simulations to be performed in limited time and on inexpensive hardware. We verified each part of the software by comparing results to analytical values where available and to micromagnetic standard problems. MUMAX3 also offers specific extensions like MFM image generation, moving simulation window, edge charge removal and material grains

Micron-sized magnetic platelets in the flux-closed vortex state are characterized by an in-plane curling magnetization and a nanometer-sized perpendicularly magnetized vortex core. Having the simplest non-trivial configuration, these objects are of general interest to micromagnetics and may offer new routes for spintronics applications. Essential progress in the understanding of nonlinear vortex dynamics was achieved when low-field core toggling by excitation of the gyrotropic eigenmode at sub-GHz frequencies was established. At frequencies more than an order of magnitude higher vortex state structures possess spin wave eigenmodes arising from the magneto-static interaction. Here we demonstrate experimentally that the unidirectional vortex core reversal process also occurs when such azimuthal modes are excited. These results are confirmed by micromagnetic simulations, which clearly show the selection rules for this novel reversal mechanism. Our analysis reveals that for spin-wave excitation the concept of a critical velocity as the switching condition has to be modified.

We present MuMax, a general-purpose micromagnetic simulation tool running on Graphical Processing Units (GPUs). MuMax is designed for high performance computations and specifically targets large simulations. In that case speedups of over a factor 100× can easily be obtained compared to the CPU-based OOMMF program developed at NIST. MuMax aims to be general and broadly applicable. It solves the classical Landau-Lifshitz equation taking into account the magnetostatic, exchange and anisotropy interactions, thermal effects and spin-transfer torque. Periodic boundary conditions can optionally be imposed. A spatial discretization using finite differences in 2 or 3 dimensions can be employed. MuMax is publicly available as open source software. It can thus be freely used and extended by community. Due to its high computational performance, MuMax should open up the possibility of running extensive simulations that would be nearly inaccessible with typical CPU-based simulators.

No abstract

No abstract