We follow a first-principles (ab initio) strategy to model magnetic materials, based on density-functional theory (DFT). The idea is to understand a material's magnetic properties in terms of the behaviour of its electrons and nuclei. The predictive power of such an approach is greatly enhanced by the fact that a minimal number of assumptions and approximations are made, and no empirical input is required.
We perform our calculations both using standard DFT packages and also with our own codes which specialize in finite-temperature properties. The codes are designed to run on high-performance parallel computers, including those maintained by the Scientific Computing Research Technology Platform at Warwick.