Research Software Engineer

Durham University - Department of Physics

The Role and Department

This role is for a Research Software Engineer within a High Performance Computing team in the Department of Physics, who will perform activities such as improving the performance of existing codes on large systems, implementing more efficient algorithms, and liaising with researchers to identify their requirements. The research areas for these codes are likely to be in the fields of cosmology, astrophysics, particle physics and nuclear physics.

The Department of Physics at Durham University is one of the leading UK Physics departments with an outstanding reputation for excellence in teaching, research and employability of our students.

The Department is committed to advancing equality and we aim to ensure that our culture is inclusive, and that our systems support flexible and family-friendly working, as recognized by our Juno Champion and Athena SWAN Silver awards. We recognise and value the benefits of diversity throughout our staff and students.

The Institute for Computational Cosmology (ICC) within the Department of Physics performs world-leading research with a high international reputation and is one of the largest groups in Europe dedicated to theoretical cosmology. The ICC operates and maintains a high performance computing (HPC) facility, COSMA, on behalf of DiRAC (www.dirac.ac.uk), a UK national facility. This service constitutes the largest academic system in the North of England and is designed to support scientific workflows which require large RAM volumes, for example cosmological simulation.

This role will be engaged with providing Research Software Engineering support for DiRAC and the ICC. Key tasks will be the improvement and development of major scientific codes, and the investigation of code performance on DiRAC services. The applicant will work with leading scientific teams to perform tasks such as code optimisation, porting to different computational architectures (which may include GPUs), and analysis of performance benchmarks. The applicant will have access to cutting edge HPC research facilities including new and novel hardware.

The applicant will also be engaged with performance benchmarking on prototype hardware to aid and guide future purchasing decisions, having a real impact on the direction of future UK HPC provision. The applicant should have a good understanding and working knowledge of C/C++ and Python. Some knowledge of Fortran would be helpful. The applicant will be using parallel scientific codes which run on multiple servers simultaneously, using middleware such as MPI. These codes will be focused on subject areas within the STFC remit, namely cosmology, particle physics and nuclear physics and the applicant will be able to work with scientists and other RSEs from across the UK.