A Research Associate position is available in the High Performance Computing Research Group at the University of Bristol. This position is associated with our Intel oneAPI Center of Excellence and the UKRI-funded ExCALIBUR Hardware and Enabling Software Benchmarking Initiative project. The oneAPI Center of Excellence builds on a history of close collaboration between Bristol University’s High Performance Computing Research Group and Intel on performance portability and advanced computer architecture research. The Center will develop best practices for performance portability at exacale using open and cross-architecture programming models. ExCALIBUR is a UK-wide program to prepare UK science codes for exascale. The Benchmarking Initiative project aims to enable benchmarking performance portability on the UK’s stable of supercomputers simple and ubiquitous.
The role will work with several collaborators from industry and academia, including Intel and NVIDIA, other oneAPI Center of Excellences, University College, London and the University of Cambridge Open Zettascale Lab, among others. The role will also provide opportunities for national and international travel, including to attend relevant conferences.
What will you be doing?
This research role will develop the state-of-art of performance portability at exascale by developing, optimising and benchmarking high performance applications on a diverse range of the latest hardware from different processor vendors. In particular, this role will develop and optimise performance portable mini-apps to target CPUs and GPUs from Intel and other vendors using open standard heterogeneous programming models. Additionally, this role will improve the tools to measure performance portability, by incorporating the performance portability methodologies developed at Bristol in collaboration with Intel into widely used benchmarking software
You should apply if
Applicants should have experience of programming, optimising and benchmarking scientific applications for heterogenous high performance supercomputers, using parallel programming models such as MPI, SYCL, OpenMP and/or CUDA.
We welcome applications from all members of our community and are particularly encouraging those from diverse groups, such as members of the LGBT+ and BAME communities, to join us.