As high-performance computing systems continue to grow in scale, improving energy efficiency without compromising scientific accuracy has become one of the major challenges facing computational science. On 27 May, CEEC hosted the second edition of our webinar on Enabling Mixed-Precision with VerifiCarlo: Sharing CEEC Experience, bringing together researchers, HPC specialists, software developers, and representatives from industry and public research organisations across Europe.
Building on the success of the original webinar in March 2025, this repeat session attracted twice as many participants, reflecting the growing interest in practical strategies for reducing the computational and energy costs of large-scale CFD applications.
During the webinar, Roman Iakymchuk (Umeå University) and Pablo de Oliveira Castro (Université Paris-Saclay UVSQ) presented the methodology developed within CEEC for introducing mixed precision into production CFD codes while maintaining numerical reliability.
Rather than simply replacing double precision with lower-precision arithmetic, the approach combines several complementary techniques to identify where reduced precision can safely be used. Using VerifiCarlo, an open-source tool built on the LLVM compiler infrastructure, researchers can analyse floating-point behaviour, detect precision-sensitive regions of code, and evaluate the numerical impact of different precision choices before modifying an application.
The speakers then demonstrated how this methodology has been applied to CEEC’s spectral element code Neko via the lightweight proxy mini-application Nekbone, showing how careful precision analysis can lead to significant improvements in both execution time and energy consumption while preserving scientific correctness. The webinar highlighted how computer arithmetic analysis, performance modelling, and expert validation together provide a systematic pathway for developing reliable mixed-precision implementations suitable for modern heterogeneous HPC systems.
Beyond the technical presentations, the webinar concluded with an active discussion on practical implementation challenges, future applications, and opportunities for collaboration. The breadth of participants—from universities and national supercomputing centres to research institutes and industrial organisations—demonstrated the wide relevance of energy-efficient numerical computing across the European HPC community.
As CEEC continues to share its experience through webinars, best practice guides, and open-source software, we hope these activities will help researchers and developers prepare their applications for future exascale systems while reducing both computational cost and energy consumption.
Further Reading
- R. Iakymchuk et al., ‘Best Practice Guide — Harvesting energy consumption on European HPC systems: Sharing Experience from the CEEC project’, Zenodo, Aug. 2024. doi: 10.5281/zenodo.13306639.
- Y. Chen, P. de O. Castro, P. Bientinesi, N. Jansson, and R. Iakymchuk, ‘Enabling Mixed-Precision in Computational Fluids Dynamics Codes’, Mar. 03, 2025, arXiv: arXiv:2503.02134. doi: 10.48550/arXiv.2503.02134.
- Y. Chen, P. de O. Castro, P. Bientinesi, and R. Iakymchuk, ‘Enabling mixed-precision with the help of tools: A Nekbone case study’, May 17, 2024, arXiv: arXiv:2405.11065. doi: 10.48550/arXiv.2405.11065.