My PhD research was focused within the field of High Performance Computing (HPC), in which I studied the performance of computational science applications and how it may be improved on modern multi-core clusters. The main focus of my research has been on the hybrid programming model, combining message passing programming (using MPI) with shared memory threading (using OpenMP).
I completed work on a hybrid MPI/OpenMP Molecular Dynamics (MD) application, the performance analysis of this code found that the hybrid model can deliver performance benefits, depending on the system interconnect used within a HPC cluster. I followed this with an analysis of a hybrid version of a real world MD application: DL_POLY. Results from both these codes have been published recently. Following this I worked towards creating a performance model describing the performance of hybrid and pure MPI MD codes.