Integrated Simulation at the Exascale: coupling, synthesis and performance
Lead Research Organisation:
University of Cambridge
Department Name: Engineering
Abstract
The arrival in the coming years of exascale computers will not just enable bigger, higher-fidelity and faster computations, but also whole new classes of simulation and modelling. It will open new frontiers in our ability to design, optimise and predict highly complex and coupled engineered and natural systems. System-level simulation of complex problems governed by multiple coupled physical processes will become possible, unlocking opportunities to create new, sophisticated engineered systems, with efficient computer simulations of interacting physical processes having the potential to greatly advance progress in high-priority areas and engineering grand challenges.
This project draws together a multidisciplinary team of leading researchers in computational science, high-performance computing, engineering and computational mathematics to create new and necessary mathematical and software tools to make stable, accurate and efficient simulation of integrated systems with coupled physical phenomena possible. It will combine rigorous mathematical analysis with cutting edge software tools to deliver new tools that will open frontiers in computing for science and engineering. The software tools will be open-source, with community building and knowledge exchange a focus throughout.
Three grand challenge problems of high social and industrial impact will direct the technical developments in this project:
- Coupled simulation of fusion modelling, which will support the virtual design and optimisation of future fusion energy systems for the electricity grid, which will have a transformative on reducing CO2 emissions;
- Carbon neutral flight, an in particular new high energy density electric propulsion systems in which the electromagnetic, thermal, mechanical and fluid process are strongly coupled; and
- Coupled simulation techniques for computing the behaviour of large virus structures.
This project draws together a multidisciplinary team of leading researchers in computational science, high-performance computing, engineering and computational mathematics to create new and necessary mathematical and software tools to make stable, accurate and efficient simulation of integrated systems with coupled physical phenomena possible. It will combine rigorous mathematical analysis with cutting edge software tools to deliver new tools that will open frontiers in computing for science and engineering. The software tools will be open-source, with community building and knowledge exchange a focus throughout.
Three grand challenge problems of high social and industrial impact will direct the technical developments in this project:
- Coupled simulation of fusion modelling, which will support the virtual design and optimisation of future fusion energy systems for the electricity grid, which will have a transformative on reducing CO2 emissions;
- Carbon neutral flight, an in particular new high energy density electric propulsion systems in which the electromagnetic, thermal, mechanical and fluid process are strongly coupled; and
- Coupled simulation techniques for computing the behaviour of large virus structures.
Organisations
- University of Cambridge (Lead Research Organisation)
- EDF Energy R&D UK Centre Limited (Project Partner)
- United Kingdom Atomic Energy Authority (Project Partner)
- Lawrence Berkeley National Laboratory (Project Partner)
- Durham University (Project Partner)
- Airbus Group Limited(Airbus Group Ltd) (Project Partner)
- Lawrence Livermore National Laboratory (Project Partner)
Publications
Dean J
(2023)
Design and analysis of an exactly divergence-free hybridised discontinuous Galerkin method for incompressible flows on meshes with quadrilateral cells
in Computer Methods in Applied Mechanics and Engineering
| Title | Software, Dataset, and Techreport: Mixed-precision finite element kernels and assembly: Rounding error analysis and hardware acceleration |
| Description | This upload contains a techreport titled "Mixed-precision finite element kernels and assembly: Rounding error analysis and hardware acceleration" together with the software (with documentation) and dataset generating the results. The software is also available on GitHub at https://github.com/croci/mpfem-paper-experiments-2024/ . The GitHub version may be updated in the future. This upload corresponds to commit number 8506dd368b84655201c8c72b1307239b9b4e43fd . See README.md file for installation instructions. The manuscript is also available on the arXiv: https://arxiv.org/abs/2410.12614. |
| Type Of Material | Database/Collection of data |
| Year Produced | 2024 |
| Provided To Others? | Yes |
| URL | https://zenodo.org/doi/10.5281/zenodo.13941628 |
| Title | Software, Dataset, and Techreport: Mixed-precision finite element kernels and assembly: Rounding error analysis and hardware acceleration |
| Description | This upload contains a techreport titled "Mixed-precision finite element kernels and assembly: Rounding error analysis and hardware acceleration" together with the software (with documentation) and dataset generating the results. The software is also available on GitHub at https://github.com/croci/mpfem-paper-experiments-2024/ . The GitHub version may be updated in the future. This upload corresponds to commit number 8506dd368b84655201c8c72b1307239b9b4e43fd . See README.md file for installation instructions. The manuscript is also available on the arXiv: https://arxiv.org/abs/2410.12614. |
| Type Of Material | Database/Collection of data |
| Year Produced | 2024 |
| Provided To Others? | Yes |
| URL | https://zenodo.org/doi/10.5281/zenodo.13941629 |
