Lattice-Boltzmann HPC
Distributed optimisations of the Lattice-Boltzmann method for computational fluid dynamics (NAG HPC Student Prize).
Development and evaluation of distributed optimisations of the Lattice-Boltzmann method for computational fluid dynamics (May 2023). Awarded the NAG HPC Student Prize for second highest accreditation.
- Serial, architecture-specific optimisations including memory allocation and vectorisation directives.
- OpenMP and MPI implementations for elastic scaling across multi-node, multi-thread systems.
- OpenCL implementation for computational offloading to GPUs.