PSyclone: a source-to-source transformation tool for Fortran code
Séminaire AMAC: EDP-AIRSEA-CVGI
11/01/2024 - 11:30 Dr. Sergi Siso (Science and Technology Facilities Council (STFC), UK) IMAG 106
PSyclone is a code-generation and transformation system designed to enable performance portability and code maintainability for weather and climate codes written in Fortran. To achieve this, it separates the scientific code, written in Fortran, from the optimization and parallelization steps, encoded as python scripts, that are needed to take advantage of each hardware platform. PSyclone is being used to modify existing directly-addressed MPI applications and offload their computations to GPUs in codes found in ocean modelling, such as NEMO, and parametrisation schemes found in atmosphere models, such as UKCA. In addition, PSyclone supports the development of Fortran-embedded DSLs, such as the PSyKAl kernel-based parallelism model that the U.K. Met Office is using to provide distributed memory parallelism for their next generation atmospheric model (LFRic).