Containerisation shows enormous potential for both simulation and analysis workflows. This project encompasses a range of mini-app tests of the applicability of containers for existing HPC workflows as well as full-on integration of significant codes within containers.
Managing large scale data systems requires metadata systems to match.
Splitting science code from the nitty gritty details of exploiting parallelism on specific computer architectures using DSLs for weather and climate : PSyclone and the new LFRic Weather and Climate Model.
Managing and analysing petascale data requires dedicated computing systems with a specialised software environment. We have provided leadership in the development of JASMIN as well as investigating real scientific workflows, developing infrastructure software, and understanding how best to deploy the necessary hardware and software infrastructure.
Two projects addressing a) optimal use of storage by developing new storage interfaces which hide storage complexity, and b) minimising data movement within simulations and in analysis workflows.
Volumes of data being generated by simulations and next generation observation platforms are huge, but not all the data being produced needs to be stored as is. We are investigating techniques to identify data that does not need to be stored (and conversely identify data that is important), and to compress data when it does need to be stored.
ACES involvement in ESIWACE includes developing new I/O sub-systems capable of handling pre-exascale data flows both at simulation time, and in the analysis environment.