Mumax3

Official Resources

• Homepage: https://mumax.github.io/
• Documentation: https://mumax.github.io/api.html
• Source Repository: https://github.com/mumax/3
• License: GNU General Public License v3.0

Overview

Mumax3 is a GPU-accelerated micromagnetic simulation program. It solves the time-dependent Landau-Lifshitz-Gilbert (LLG) equation using finite difference discretization. It is designed to be highly efficient, running exclusively on NVIDIA GPUs, and provides a rich scripting interface (Go-like syntax) for defining complex geometries, time-dependent fields, and various magnetic interactions.

Scientific domain: Micromagnetics, spintronics, magnetic dynamics
Target user community: Magnetism researchers, spintronics engineers, device physicists

Theoretical Methods

• Landau-Lifshitz-Gilbert (LLG) equation
• Finite Difference Method (FDM)
• Exchange interaction (Heisenberg)
• Dipole-Dipole interaction (Demagnetization)
• Magnetic Anisotropy (Uniaxial, Cubic)
• Dzyaloshinskii-Moriya Interaction (DMI) - Bulk and Interfacial
• Spin-Transfer Torque (Zhang-Li, Slonczewski)
• Thermal fluctuations (Langevin dynamics)

Capabilities (CRITICAL)

• GPU-accelerated micromagnetic simulations (very fast)
• Simulation of domain walls, skyrmions, vortices
• Spin-transfer torque and spin-orbit torque
• Time-dependent excitation (RF fields, pulses)
• Complex geometries via constructive solid geometry (CSG) or image masks
• Voronoi tessellation for grains
• Periodic boundary conditions

Sources: Mumax3 website, AIP Advances 4, 107133 (2014)

Inputs & Outputs

• Input formats: .mx3 script files, image files (png) for masks
• Output data types: .ovf (vector fields), .txt (tables), .png (snapshots)

Interfaces & Ecosystem

• Go: Scripting language based on Go
• Python: mumax3-python wrapper available
• Visualization: Output compatible with ParaView, Muview, OOMMF tools
• Web Interface: Mumax3-server

Workflow and Usage

1. Write .mx3 script: Define geometry, material parameters, and simulation protocol.
2. Run simulation: mumax3 script.mx3
3. Convert output: mumax3-convert to standard formats (OVF/VTK).
4. Analyze tables and visualize vector fields.

Performance Characteristics

• Optimized for NVIDIA CUDA
• Speedup of 10x-100x compared to CPU codes (like OOMMF)
• Performance scales with number of CUDA cores

Application Areas

• Magnetic storage devices (MRAM, HDD)
• Skyrmionics and topological magnetism
• Magnonics (spin waves)
• Nanomagnetism and logic devices

Community and Support

• Open-source (GPL v3)
• Developed by DyNaMat group (Ghent University)
• Active Google Group for support
• Large user community in micromagnetics

Verification & Sources

Primary sources:

1. Homepage: https://mumax.github.io/
2. GitHub: https://github.com/mumax/3
3. Publication: A. Vansteenkiste et al., AIP Advances 4, 107133 (2014)

Confidence: VERIFIED

Verification status: ✅ VERIFIED

• Website: ACTIVE
• Documentation: COMPREHENSIVE
• Source: OPEN (GitHub)
• Development: ACTIVE/MAINTAINED
• Applications: Micromagnetics, GPU acceleration, LLG