SC22 Proceedings

The International Conference for High Performance Computing, Networking, Storage, and Analysis

Birds of a Feather Archive

ExCALIBUR : Exploring Successes, Collaborations and Community Building Opportunities with the UK Exascale Effort

Authors: Nick Brown (Edinburgh Parallel Computing Centre (EPCC), University of Edinburgh), Stephen Longshaw (STFC Daresbury Laboratory, England; UK Research and Innovation), Chris Maynard (University of Reading; Met Office, UK)

Abstract: The Exascale Computing ALgorithms & Infrastructures Benefiting UK Research (ExCALIBUR) program is a research effort aiming to enable exploitation of future exascale supercomputers by the next generation of high-performance simulation software. Funded by the UK government, and running between 2019 and 2025, the program focuses on targeting high priority codes, algorithms, and techniques to meet the demands of computational scientists and engineers. Currently at the mid-point of the program, in this BoF we will highlight some of the activities and successes to date, as well as explore opportunities for collaborating more widely with global exascale computing research activities and programs.

Long Description: The UK has a long history of investing in supercomputers, for instance the current UK national supercomputer, ARCHER2, which is a 750,000 core Cray EX and there are plans to obtain an exascale system by middle of the decade. However, as these machines become larger and more complex, a key exploitation challenge lies in the software programming and algorithmic aspects.

With this in mind, in 2019 the UK government funded the Exascale Computing ALgorithms & Infrastructures Benefiting UK Research (ExCALIBUR) programme. Running until 2025 and focused heavily, although not exclusively, around weather and climate (UK Met Office) and fusion energy (UK Atomic Energy Authority) workloads, this programme aims to further develop and refine the major software and algorithmic building blocks that will be required for exascale supercomputing.

The ExCALIBUR programme itself comprises four major themes; use-cases, exascale techniques and technologies, next-generation computing hardware, and investment in people. Initially starting with projects aimed at identifying concrete supercomputing use-cases and requirements across weather and climate, fusion energy, and other key scientific and engineering domains, the lessons learnt from these then fed into calls for exascale techniques and technology activities. These aim to work with the use-cases in supporting their use of large-scale HPC, and have ranged across areas including domain specific languages, quantum computing, machine learning for computation, exascale IO, workflows, simulation code coupling, task-based parallelism, and parallel in time methods.

Concurrently, activities to enable access to next-generation computing architectures for ExCALIBUR projects and the HPC community more widely have been undertaken. This involves providing testbeds such as FPGAs, RISC-V, Graphcore, ARM, and Cerebras CS-2, as well as benchmarking activities for evaluation. Testbeds work closely with vendors as technology partners, looking to support co-design and impact vendor roadmaps where possible. Lastly, ExCALIBUR also funded projects aimed at supporting, encouraging, and training the current and next-generation of HPC scientific software developers (Research Software Engineers). This also involves understanding how best to ensure that RSE careers are fulfilling and the RSEs themselves valued, addressing the skills shortage that we currently face in HPC.

We are now at the mid-point of ExCALIBUR, and there have been numerous successes across all themes and activities, but it is also important to look outwards and drive connections more widely with similar efforts and programmes going on elsewhere in the world. We are keen to share technologies and techniques, as well as more concrete aspects such as physical access to the ExCALIBUR testbeds and RSE training materials that we believe could benefit, for instance, US exascale efforts such as ECP and Frontier.

Consequently, a BoF at supercomputing will be a very valuable opportunity to widely share with the SC community what we are doing at the programme level, explore how we might work more closely together in future, and global community building. In this session we will aim to disseminate with the community what ExCALIBUR projects are currently being undertaken, what has been achieved so far, and explore opportunities for collaboration and sharing with other programmes, groups, and individuals.


Back to Birds of a Feather Archive Listing