CEEC Progress

For centuries, scientific discovery has relied on experiments to understand new theoretical models. However, the advent of powerful supercomputers and efficient numerical methods has made it possible to perform simulations instead. Computational Fluid Dynamics (CFD) provides a powerful tool for qualitative and quantitative analysis of fluid flow by solving a set of equations on a computer. This enables scientists and engineers to perform experiments in a virtual laboratory and study a variety of fluid flows encountered in everyday life. That is why CFD is used today in various disciplines, from basic scientific research to engineering applications. However, unlike many other scientific fields, there is virtually no upper limit to the problem size in fluid dynamics, thus making detailed simulations exceedingly expensive to perform. The main goal of the Centre of Excellence in Exascale CFD (CEEC) is to address extreme-scale computing challenges to make it possible to use large-scale, accurate and cost-efficient high-fidelity CFD simulations for both academic and industrial applications.

At the end of 2024, CEEC reached the halfway mark of the project. During the first two years of CEEC, we completed the initial requirements phase, defining lighthouse use cases, which are a central part of CEEC around which the rest of the centre of excellence is organised. They provide the necessary requirements for our algorithmic work packages. We have also been starting to see what the end of the project’s development phase might look like, where CEEC has made significant progress in ensuring exascale readiness of the codes supporting the lighthouse cases. All of the core codes have now been ported and tuned for accelerators and are running on all of the currently installed EuroHPC Joint Undertaking (EuroHPC JU) pre-exascale machines; some are even participating in the exascale pilot for the first European exascale machine, JUPITER.

A large part of the current development phase of CEEC is dedicated to improving algorithms for efficient exploitation of exascale architectures and also to enhancing the techniques and technologies required for our lighthouse cases. At the halfway mark, these algorithmic developments are being integrated into our core codes to support the execution of the lighthouse codes. For techniques and technologies, we see a similar trend: lighthouse cases are being coupled with machine learning interfaces, workflow environments, and enhanced visualisation and data management tools.

CEEC is now working on the second half of the project goals. We are entering the demonstration phase and have been working on finalising the integration and implementation of our work with the lighthouse cases. We will be moving forward to demonstrating the capabilities on available pre-exascale and exascale systems.

The work done by CEEC so far has been highly successful. In particular, some of the early developments in CEEC formed a key part of work that reached the final of the Gordon Bell Prize in 2023. This prize is awarded each year to recognise outstanding achievement in high-performance computing (HPC), such as innovation in applying HPC to applications in important engineering science, and large-scale data analytics problems.