SysGenX: Composable software generation for system-level simulation at Exascale
Lead Research Organisation:
UNIVERSITY COLLEGE LONDON
Department Name: Mathematics
Abstract
Abstracts are not currently available in GtR for all funded research. This is normally because the abstract was not required at the time of proposal submission, but may be because it included sensitive information such as personal details.
Organisations
People |
ORCID iD |
| Timo Betcke (Principal Investigator) | |
| Kai Luo (Co-Investigator) |
Publications
Daversin-Catty C
(2024)
Finite Element Software and Performance for Network Models with Multipliers
Daversin-Catty C
(2024)
Finite Element Software and Performance for Network Models with Multipliers
Fierro-Piccardo I
(2023)
An OSRC Preconditioner for the EFIE
in IEEE Transactions on Antennas and Propagation
Kailasa S
(2022)
PyExaFMM: An Exercise in Designing High-Performance Software With Python and Numba
in Computing in Science & Engineering
Lei T
(2023)
Study of CO 2 desublimation during cryogenic carbon capture using the lattice Boltzmann method
in Journal of Fluid Mechanics
Lei T
(2024)
Pore-scale study of CO 2 desublimation in a contact liquid
in Carbon Capture Science & Technology
Lei T
(2024)
Insight into discharge of non-aqueous Li-O 2 battery using a three-dimensional electrochemical lattice Boltzmann model
in Chemical Engineering Journal
Liu Z
(2024)
Microcontinuum approach to multiscale modeling of multiphase reactive flow during mineral dissolution
in Physical Review Fluids
Mao Q
(2023)
Classical and reactive molecular dynamics: Principles and applications in combustion and energy systems
in Progress in Energy and Combustion Science
| Description | We have developed and successfully tested on ARCHER2, the UK's national supercomputing infrastructure, a novel fast multipole library for computing interactions in applications such as electrostatics, astronomy, electromagnetic, and others. The library has been developed in Rust, which is a new modern programming language, and is to our knowledge the first such HPC library in Rust. Scaling experiments have been undertaken up to one billion particles on 128 nodes. We are currently integrating the library in a wider framework for the scalable evaluation of Galerkin integral operators on supercomputers and eventual coupling with FEniCS for extreme scale FEM/BEM simulations. As part of this effort we are building a whole ecosystem of libraries in Rust, including distributed linear algebra, grid management, fast evaluation of Green's functions, and others. These libraries are highly composable and designed as open-source projects to benefit the wider community. |
| Exploitation Route | The libraries demonstrate that a viable HPC ecosystem can be established in Rust that is highly competitive with legacy C/C++ HPC libraries. The outcomes are well maintained open-source software that can easily be integrated in other projects. We have also started development of interfaces to C/C++ and Python to make use outside Rust easier. |
| Sectors | Aerospace Defence and Marine Digital/Communication/Information Technologies (including Software) |