I am a member of the School’s Numerical Linear Algebra group.
Modern parallel computers are increasingly likely to be heterogeneous, comprising, for example, multiple powerful accelerators, such as GPUs, in addition to a host of multicore CPUs. But to exploit the awesome potential of this new landscape, we must ensure that computational jobs are executed by the processing resources which are best suited to handle them. My research is focused on developing cheap and effective heuristics for such scheduling problems.
- Numerical analysis.
- Machine learning.
- Parallel/high-performance computing (HPC).
During the 2017/18 academic year, I was the Treasurer of the University of Manchester SIAM-IMA Student Chapter.