One of the services will also be supported by an additional £3 million investment from the Science and Technology Facilities Council, Medical Research Council, Health Data Research UK and UK Atomic Energy Authority (Culham Centre for Fusion Energy).
The Tier-2 supercomputing services will provide researchers with invaluable access to powerful systems to support ground-breaking work in areas ranging from Artificial Intelligence (AI), energy storage and supply and therapeutic drug design.
The funding announced today will not only provide new state-of-the-art computing hardware across a wide range of different technologies, but will support the development of research computing skills, including boosting the careers of Research Software Engineers, across the UK.
EPSRC Executive Chair Professor Dame Lynn Gladden said: “Computation is becoming an ever-more important scientific tool, be it for analysing large data sets generated from experimental work or modelling situations which can’t be replicated in experiments.
“The High Performance Computing services announced today will give researchers access to the tools they need to make breakthroughs in a wide range of fields that impact on how we live our lives.
“These include heterogeneous catalysis – modelling chemical processes which contribute to the production of items used in everyday life - understanding the performance of materials for better batteries for electric vehicles and other energy storage applications, and using advanced computational drug design for therapeutics targeting a large variety of health conditions.”
Summaries of the services
Cambridge Service for Data Driven Discovery (CSD3) – A National Data Intensive Science Cloud for Converged Simulation, AI and Analytics
Led by: University of Cambridge
Partners: Cardiff University, King’s College London, Universities of Edinburgh, Leicester, Oxford and Southampton
EPSRC support: £4 million
The Cambridge Service for Data Driven Discovery is a £10M investment, creating one of the most powerful academic supercomputers in the UK specially designed to combine large scale data intensive simulation and AI science within a single computer system.
This breakthrough supercomputer was developed via a unique Co-design partnership between the University of Cambridge, Dell, Intel, NVIDIA, Mellanox Technologies and StackHPC a leading UK SME developing ground breaking HPC system software. It will accelerate research across a wide range of engineering and physical science themes, including materials science and computational chemistry, health informatics, medical imaging and bio-simulation, and AI and machine learning. The proposal has stimulated significant industrial investment from Dell, Intel and Cambridge to form the Cambridge Open Exascale Lab, providing a large critical mass of people and technology focused on driving UK competitiveness in the rapidly emerging exascale computing landscape.
The service will also be supported by an additional £3 million investment from the Science and Technology Facilities Council, Medical Research Council, Health Data Research UK and UK Atomic Energy Authority (Culham Centre for Fusion Energy).
The Materials and Molecular Modelling Hub
Led by: UCL
Partners: Queen Mary University of London, Queen’s University Belfast, Brunel University, Imperial College London, King’s College London, Universities of Cambridge, Lincoln, Kent, Reading, Southampton and York
EPSRC support: £4.5 million
Materials are at the heart of almost every modern technology, including energy generation, storage and supply, transportation, electronic devices, defence and security, healthcare, and the environment. It is materials that place practical limits on efficiency, reliability and cost. The MMM Hub provides high performance computing capacity for researchers to carry out ground-breaking research on the properties of new and existing materials, and this funding will build on the hub’s capability.
These include understanding and preventing surface degradation, such as corrosion and wear, on a range of different materials; researching how changes to the recycling of metals can reduce the environmental damage caused by metal extraction; and developing the next generation of materials for solar energy generation.
GW4 Tier-2 HPC Centre for Advanced Architectures (Isambard 2)
Led by: The GW4 Alliance of the Universities of Bath, Bristol, Cardiff and Exeter, and hosted by the Met Office
EPSRC support: £4.1 million
The Isambard 2 service will use the very latest technology from the UK-based Arm Holdings to provide scientists with a world-class High Performance Computing service. The same technology is expected to be used in some of the first supercomputers capable of a billion billion calculations per second, called Exascale supercomputers. This technological step change will require scientists to adapt their codes in order to run as quickly and efficiently as possible and thus accelerate scientific discovery. Isambard 2 will succeed the current Isambard system and will help enable British researchers to prepare their codes for the widespread use of Exascale systems.
Isambard has already been used to investigate potential drugs to treat osteoporosis and simulate Parkinson's disease at the molecular level. Isambard 2 will enable researchers to expand this further with the potential for scientific breakthroughs.
Kelvin-2
Led by: Queen’s University Belfast and Ulster University
EPSRC support: £2.1 million
The Kelvin-2 service will provide access to an enhanced computing facility focused on artificial intelligence-based research. The project will initially focus on accelerating research in six specialist areas which are economically and socially important to the UK. This includes neurotechnology and computational neuroscience, including work on brain-computer interfaces and heterogeneous catalysis, such as modelling chemical processes which contribute to the production of items used in everyday life.
The areas also include innovative drug delivery for improving drug-based therapies and for use in diagnostics, as well as a focus on precision medicine where automated tools will be created to analyse data and identify indicators for health conditions. There will also be a focus on food fingerprinting, including techniques for detecting chemical contaminants in food; and hydrogen deflagration to assist with developing accident prevention and mitigation for hydrogen tanks.
JADE: Joint Academic Data Science Endeavour – 2 (JADE 2)
Led by: University of Oxford
EPSRC support: £5.5 million
Partners: Universities of Bath, Bristol, Cambridge, Exeter, Lancaster, Leeds, Loughborough, Sheffield, Southampton, Surrey, Sussex, Warwick and York, Queen Mary University of London, King’s College London, Imperial College London, UCL, Newcastle University, The Alan Turing Institute, Hartree Centre.
The JADE 2 service, hosted at STFC’s Hartree Centre, will be a unique national resource providing a state-of-the-art GPU (Graphics Processing Unit) computing facility for research into Artificial Intelligence (AI)/Machine Learning and Molecular Dynamics, a computer simulation method for analysing the physical movements of particles that make up molecules.
The AI techniques developed using the service will have impact in a range of sectors including financial services, manufacturing, energy and healthcare. The Molecular Dynamics research conducted on JADE 2 will advance understanding of the structure and function of large biological molecules, many of which are targets for therapeutic agents for a large variety of health conditions. The service will provide a valuable computing resource to the new UKRI Artificial Intelligence Centres for Doctoral training, and thus provide a critical computational capacity needed to develop the next generation of experts in AI.
Cirrus Phase II: Preparing for Heterogeneity at Exascale
Led by: The Edinburgh Parallel Computing Centre at the University of Edinburgh
EPSRC support: £3.5 million
Cirrus Phase II will expand the capabilities of the Cirrus service by adding specialized GPUs to the current system. GPUs are commonly used as graphics/video cards in mobile phones, personal computers and games consoles. However, specialized GPUs can be also be used in supercomputers as accelerators enabling them to run numerical calculations more quickly. The technology used in Cirrus is expected to be used in some of the first Exascale supercomputers and will allow scientists to test and adapt their code for modelling and simulation to be ready to advance discovery and innovation as soon as Exascale systems become available. It will help to ensure the UK has a supply of individuals trained with these specialised skills and could lead to far more rapid and detailed discoveries in new areas and the projects Cirrus has already supported, such as modelling protein shape for better drug design and simulating tidal flows to optimise turbine installations and their effects on sea beds. The new GPUs will also provide a high-performance platform for AI training and research, a critical and rapidly growing area.
Northern Intensive Computing Environment (NICE)
Led by: Durham University as part of the N8 Research Partnership (Durham and Newcastle Universities, Universities of Lancaster, Leeds, Liverpool, Manchester, Sheffield and York)
EPSRC support: £3.1 million
The new NICE service will use the same technology as that used in the current leading supercomputers in the world, extending the capability of accelerated computing. The technology has been chosen with the aim of combining experimental, modelling and machine learning approaches, bringing these communities together to address new challenges.
This will mean that machine learning can be used alongside modelling and simulation to better understand the vast data sets now being generated by experimentalists through, for example, national facilities such as Diamond and the Henry Royce Institute and international facilities such as the European Synchrotron Radiation Facility. This approach will enable scientists to, for example, advance the imaging techniques necessary to produce the next generation of X-ray instruments and develop future students working with deep learning techniques at the interface of algorithms and High Performance Computing.