Towards global km-scale weather and climate simulation on next-generation supercomputers (Oliver Fuhrer, Météo Suisse)

21.12.2017 16:15

Should global warming occur at the upper end of the range of current projections, the local impacts of unmitigated climate change would be dramatic. Despite significant progress, climate projections with so called global climate models still entail large uncertainties. There is overwhelming evidence, that one of the key reasons for this large uncertainty is the representation of clouds using horizontal grid resolutions on the order of 50-200 km. There is hope that next-generations HPC systems will allow for km-scale global climate simulations that would significantly decrease these uncertainties.

But, the weather and climate community is struggling to keep up with the fast-paced developments as HPC system designs strive towards the Exascale. Only few production-ready community codes are able to leverage emerging hardware architectures. This has become known as the software productivity gap. While programming models - such as OpenACC - are available to target these hardware architectures, they further increase the complexity of the code and thus exacerbate the problem. This talk describes a co-design effort for a regional weather and climate model that presents an alternative pathway.

How far away are we today from the target of reaching km-scale global weather and climate simulations? Using the regional weather and climate model, we assess what can be achieved today using near-global simulations on Europe's largest supercomputer Piz Daint and discuss challenges and issues from a numerical and high-performance computing perspective for achieving this target for production simulations.

Lieu

salle 623, Séminaire d'analyse numérique

entrée libre