Many research problems in science and engineering rely on the solution of partial differential equations (PDEs) on a large scale and at a high resolution; one example is numerical weather prediction (NWP). This requires the use of modern high performance computing (HPC) technology and corresponding modern discretisation techniques, specifically the Discontinuous Galerkin (DG) method. This type of discretisation has been studied since the 1960's, but is becoming more popular for applications due to properties that make them efficient on modern HPC architectures.

In this talk I will introduce the DG method and work though an example. I will also highlight some of the desirable properties of the method and why you may want to consider using DG over other numerical methods for solving PDEs.