Computational biomedicine offers many avenues for taking full advantage of emerging exascale computing resources and, as such, will provide a wealth of benefits as a use-case within the wider ExCALIBUR initiative. These benefits will be realised not just via the medical problems we elucidate but also through the technical developments we implement to enhance the underlying algorithmic performance and workflows supporting their deployment. Without the technical capacity to effectively utilise resources at such unprecedented scale - either in large monolithic simulations spread over the equivalent of many hundreds of thousands of cores, in coupled code settings, or being launched as massive sets of tasks to enhance drug discovery or probe a human population - the advances in hardware performance and scale cannot be fully capitalised on. Our project will seek to identify solutions to these challenges and communicate them throughout the ExCALIBUR community, bringing the field of computational biomedicine and its community of practitioners to join those disciplines that make regular use of high-performance computing and are also seeking to reach the exascale. In this project, we will be deploying applications in three key areas of computational biomedicine: molecular medicine, vascular modelling and cardiac simulation. This scope and diversity of our use cases mean that we shall appeal strongly to the biomedical community at large. We shall demonstrate how to develop and deploy applications on emerging exascale machines to achieve increasingly high-fidelity descriptions of the human body in health and disease. In the field of molecular modelling, we shall develop and deploy complex workflows built from a combination of machine learning and physics-based methods to accelerate the preclinical drug discovery pipeline and for personalised drug treatment. These methods will enable us to develop highly selective small molecule therapeutics for cell surface receptors that mediate key physiological responses. Our vascular studies will utilise a combination of 1D, 3D models and machine learning to examine blood flow through complex, personalised arterial and venous structures. We will seek to utilise these in the identification of risk factors in clinical applications such as aneurysm rupture and for the management of ischaemic stroke. Within the cardiac simulation domain, a new GPU accelerated code will be utilised to perform multiscale cardiac electrophysiology simulations. By running large populations based on large clinical datasets such as UK Biobank, we can identify individual at elevated risk of various forms of heart disease. Coupling heart models to simulations of vascular blood flow will allow us to assess how problems which arise in one part of the body (such as the heart) can cause pathologies on remote regions. This exchange of knowledge will form a key component of CompBioMedX. Through this focussed effort, we will engage with the broader ExCALIBUR initiative to ensure that we take advantage of the efforts already underway within the community and in return reciprocate through the advances made with our use case. Many biomedical experts remain unfamiliar with high-performance computing and need to be better informed of its advantages and capabilities. We shall engage pro-actively with medical students early in their career to illustrate the benefits of using modelling and supercomputers and encourage them to exploit them in their own medical research. We shall engage in a similar manner with undergraduate biosciences students to establish a culture and practice of using computational methods to inform the experimental work underpinning the basic science that is the first step in the translational pathway from bench to bedside.

This proposal brings together 19 universities, including 12 out of 16 newly established UKRI CDTs in Artificial Intelligence. Led by the University of Oxford, with support from the Alan Turing Institute (Turing), Bath, Bristol, Cambridge, Exeter, Imperial, KCL, Leeds, Loughborough, Newcastle, QMUL, Sheffield, Southampton, Surrey, Sussex, UCL, Warwick and York, our proposal aims to build on the success of the JADE Tier 2 facility. The current JADE facility represents a unique national resource providing state of the art GPU computing facilities to world leading experts in the areas of Artificial Intelligence/Machine Learning (AI/ML) and molecular dynamics (MD) research. In addition to providing a leading compute resource, the JADE facility has also provided a nucleus around which a national consortium of AI researchers has formed, making it the de facto national compute facility for AI research. By providing a much-needed shared resource to these communities, JADE has also delivered an outstanding level of world leading science, evidenced in the twenty two pages of preliminary case studies submitted to EPSRC on 11/09/18. JADE2 will build upon these successes by providing increased computational capabilities to these communities and delivering a stronger, more robust service to address the lessons learned from the initial service. The architecture for JADE2 will be a similar to that of JADE, based on NVIDIA's DGX platform. JADE is formed from 22x DGX1V nodes. JADE2 will be over twice the size of JADE and employ the more cost effective DGX1 Max Q platform. Differences between Max Q and the premium DGX1V are centred on on a slightly reduced bandwidth to GPU memory and lower peak compute performance. Tests of relevant codes on these platforms show that, for AI/ML and Molecular Dynamics, Max Q achieves at least 3/4 performance, using 2/3 the power for 1/2 of the price. The system will be run as a national facility, providing free access to all academic users through a lightweight Resource Allocation Panel (RAP). HECBioSim will run the RAP for MD users, ATI will run the corresponding RAP for AI/ML users.