Simulation has become a key tool in research and development of magnetic nanostructure systems, including data storage and sensing. The simulation of magnetisation at the nanoscale poses computational challenges as the relevant time-dependent partial differential equations are non-linear and stiff. The formulation of the micromagnetic model requires a fine discretisation, resulting in many degrees of freedom. New design proposals can be explored virtually, and help making the device design process more cost-effective and efficient but require simulation of large structures.

The workhorse of micromagnetic modelling is a finite-difference based software that executes on single and multi-core CPUs (OOMMF [1]). The availability of large computational power in GPUs at low costs has led to the development of further micromagnetic codes that have been written especially for (NVidia) GPUs, such as the open source codes Mumax3 and Micromagnum, and a few other codes that are not publicly available.

In this project, we will explore feasibility of developing similar high performance computing code that can flexibly execute on several computational backends, such as GPUs, Intel Phi and multi-core/multi-CPU systems. We are aiming for an implementation where the equations to solve are coded once but the translation into backend-specific code happens automatically. In the long run, this will lead to more sustainable code that can be re-used despite rapidly changing hardware.

We will use a design where the end user of the application uses a Python interface, similar to the Nmag [2] simulation design.

Eventually, we will apply the prototype simulation tool to state-of-the-art simulation challenges from the micromagnetic community to advance science and evaluate performance and usefulness of the tools, together with our national and international collaborators. Topics of particular interest in the micromagnetic field include Skyrmionics and Magnonics, and the use of these emerging fields in new devices.

[1] http://math.nist.gov/oommf/

[2] http://nmag.soton.ac.uk

If you wish to discuss any details of the project informally, please contact Dr Sam Sinayoko, Fluid Dynamics and Acoustics research group, Email: s.sinayoko<στο>soton.ac.uk, Tel: +44 (0) 2380 593082.

Funding information: This project is in competition with others for the associated funding. The funding covers EU/UK fees and stipend. 

This project is run through participation in the EPSRC Centre for Doctoral Training in Next Generation Computational Modelling (http://ngcm.soton.ac.uk).