Spirit Code Structure
This Page explains the code structure of the Spirit code, it's namespaces, modules and approaches.
Modules
Core
Data
Engine
Utility
API
GL
This library provides OpenGL functions for visualisations of vector fields (specialized to spins)
QT UI
This User Interface, built with QT5, provides a live visualisation, the ability to start and stop simulations and to set parameters.
Python UI
This User Interface is a convenient way of using the core library in an automated way or on systems where a Graphical User Interface is not available. An interface layer has been written to provide python functions which resemble those of the core API.
Web UI
This User Interface, written in JavaScript, CSS and HTML, has been created in order to be able to showcase the code, e.g. in presentations or on mobile devices. For this interface, no installation is required and most non-mobile browsers and a few mobile ones will work.
The Web UI may also be used for educational purposes, e.g. to study spin spirals or antiferromagnets in lower level university courses. For this, some further development of the website around this will be necessary. It might even be used to teach school classes some basic magnetism.
The Engine
Method
Run logic.
- LLG
- GNEB
- MMF
Optimizer
Iterates on Spin Systems to move them according to an applied force (independent of if it is physical or virtual).
- Semi-Implicit B (SIB)
- Heun
- Conjugate Gradient (CG)
- Quick-Min (QM)
Hamiltonian
Calculates energy and effective field which are needed by the optimizer and the force calculation. There are two cases currently implemented:
- isotropic interactions, formulated in shells of neighbours
- anisotropic interactions, formulated pair-wise
Modularity
The Method, the Optimizer (currently not yet the Hamiltonian) can be interchanged flexibly at run time. It is a goal of the developers to be able to change to a new Geometry at runtime.
This modular structure makes it very easy to extend the code and add new algorithms or physical models. In order to implement a new model, one might only need to write a new Hamiltonian, which provides a way to calculate Energies, Energy Gradients and Hessian matrices.