Back

Minisymposium Presentation

Can we Build Composable Atmospheric Models Without Sacrificing Performance?

Wednesday, June 5, 2024
13:00
-
13:30
CEST
Climate, Weather and Earth Sciences
Climate, Weather and Earth Sciences
Climate, Weather and Earth Sciences
Chemistry and Materials
Chemistry and Materials
Chemistry and Materials
Computer Science and Applied Mathematics
Computer Science and Applied Mathematics
Computer Science and Applied Mathematics
Humanities and Social Sciences
Humanities and Social Sciences
Humanities and Social Sciences
Engineering
Engineering
Engineering
Life Sciences
Life Sciences
Life Sciences
Physics
Physics
Physics

Presenter

Stefano
Ubbiali
-
ETH Zurich

After completing a double degree master program in computational science at Politecnico di Milano and EPF Lausanne, I got a PhD in applied math and computational physics from ETH Zurich. I am currently a postdoctoral researcher in the Institute for Atmospheric and Climate Science at ETH Zurich, working on the development of a performance-portable atmospheric model for all-scale predictions.

Description

Atmospheric models consist of a dynamical core – integrating the equations of motion on a computational mesh – and physical parameterizations – taking into account the bulk effect of subgrid-scale phenomena (e.g. radiative heat transfer, microphysics, turbulence). For ease of software development, dynamical cores and physics packages have historically been written in isolation, leading to model components based on inconsistent assumptions and featuring incompatible structures. We present recent efforts to devise model components with a common and expressive interface, which can be more easily transferred between models and favor modular code designs. We discuss the challenges and benefits associated with this approach, both from a software engineering perspective (e.g. maintainability, reusability, interoperability, readability) and a scientific point of view (e.g. process coupling). Moreover, we address the integration of modern HPC tools (e.g. domain-specific languages) into composable code architectures and discuss potential impacts on performance, as compared to a monolithic code design.

Authors