magnum.np -- A PyTorch based GPU enhanced Finite Difference Micromagnetic Simulation Framework for High Level Development and Inverse Design

TL;DR

magnum.np is a PyTorch-based micromagnetic simulation framework that offers high extensibility, efficient GPU execution, and facilitates inverse problem solving through autograd, advancing simulation development and inverse design.

Contribution

It introduces a novel micromagnetic simulation library built on PyTorch, enabling high-level development, GPU acceleration, and inverse problem capabilities.

Findings

Achieves competitive performance with existing micromagnetic codes.

Enables rapid development of new algorithms and functionalities.

Supports inverse problem solving via autograd.

Abstract

magnum.np is a micromagnetic finite-difference library completely based on the tensor library PyTorch. The use of such a high level library leads to a highly maintainable and extensible code base which is the ideal candidate for the investigation of novel algorithms and modeling approaches. On the other hand magnum.np benefits from the devices abstraction and optimizations of PyTorch enabling the efficient execution of micromagnetic simulations on a number of computational platforms including GPU and potentially TPU systems. We demonstrate a competitive performance to state-of-the art micromagnetic codes such a mumax3 and show how our code enables the rapid implementation of new functionality. Furthermore, handling inverse problems becomes possible by using PyTorch's autograd feature.