The International Conference for High Performance Computing, Networking, Storage and Analysis
Rolls-Royce Hydra CFD Code on GPUs Using OP2 Abstraction.
Authors: Istvan Z. Reguly (University of Oxford), Gihan R. Mudalige (University of Oxford), Carlo Bertolli (IBM Corporation), Michael B. Giles (University of Oxford), Adam Betts (Imperial College London), Paul H. J. Kelly (Imperial College London), David Radford (Rolls-Royce plc)
Abstract: Hydra is an industrial CFD application used for the design of turbomachinery, now automatically accelerated by GPUs through the OP2 domain specific "active library" for unstructured grid algorithms. From the high-level definition, either CPU or GPU code is generated, applying optimisations such as conversion to Structure-of-Arrays, use of the read-only cache or the tuning of block sizes automatically. A single GPU is over 2 times faster than the original on a server-class CPU, we demonstrate excellent strong and weak scaling, evaluated up to 4096 CPU cores or 16 GPUs.