Cambridge-1, the Nvidia-backed supercomputer, is now up and running, processing pharmaceutical and healthcare-related data and workloads to accelerate the pace of the UK’s genome sequencing, drug discovery and disease research.
Based on the Nvidia DGX SuperPod system, Cambridge-1 is reportedly the UK’s most powerful supercomputer, with the capacity to deliver 400 petaflops of artificial intelligence (AI) performance and 8 petaflops of Linpack performance.
The £50m supercomputer is powered by 80 Nvidia DGX A100 systems linked together by Nvidia Mellanox InfiniBand networking that will enable research teams to run AI training, inference and data science workloads at scale. It is also 100% renewably powered.
Nvidia claims the performance of Cambridge-1 makes it the 29th most powerful supercomputer in the Top500 list of the most powerful supercomputers, as well as being one of the world’s most energy-efficient high-performance computing (HPC) environments, with a top three ranking in the Green500 list.
Speaking at a press pre-brief ahead of the Cambridge-1 supercomputer launch event, David Hogan, vice-president of enterprise at Nvidia, said the project was indicative of how the company’s ability to serve the needs of users in the healthcare space has evolved over the course of the past decade.
“We started off [a decade ago] working in radiology and medical imaging, and then progressed into areas such as genomics. With the advent of AI, we’re working in the field of drug discovery, looking at the ability to identify targets and compounds, and more recently working with things like natural language processing to be able to process and analyse both research and clinical data to enhance the diagnostic process,” said Hogan.
User engagement in Cambridge-1 project
It is hoped the Cambridge-1 setup, which was first announced in October 2020, will enable breakthrough clinical and life sciences research into a number of health conditions, including dementia and multiple sclerosis, and also accelerate the pace of genomic sequencing and drug discovery for the project’s founding partners.
These include pharmaceutical giants AstraZeneca and GlaxoSmithKline (GSK), DNA sequencing firm Oxford Nanopore Technologies, King’s College London university and the Guy’s and St Thomas’ NHS Foundation Trust.
AstraZeneca is using the setup to create a transformer-based generative AI model for chemical structures. It is also working with Nvidia on a different project involving Cambridge-1 that will focus on the use of AI in digital pathology so that tissue sample slide images can be analysed more quickly to speed up the insights gleaned during drug response research.
GSK, meanwhile, is using Cambridge-1 to develop medicines it claims are twice as likely to be brought to market by drawing on the system’s AI and machine learning (ML) capabilities.
“Advanced technologies are core to GSK’s research and development approach and help to unlock the potential of large, complex data through predictive modelling at new levels of speed, precision and scale,” said Kim Branson, senior vice-president and head of AI-ML at GSK.
“We are pleased to have the opportunity to partner with Nvidia to deliver on GSK’s drug discovery ambition and contribute to the UK’s rich life sciences ecosystem – both aims that have patient benefit at the centre.”
Kings’s College London and Guy’s and St Thomas’ NHS Foundation Trust are using Cambridge-1 to teach AI models to generate synthetic brain images by training them on tens of thousands of MRI brain scans.
“The ultimate goal is to use this synthetic data model to gain a better understanding of diseases like dementia, stroke, brain cancer and multiple sclerosis, and enable earlier diagnosis and treatment,” the organisations said in a statement.
“As this AI synthetic brain model can generate an infinite amount of never-seen brain images with chosen characteristics, it will allow a better and more nuanced understanding of what diseases look like, possibly enabling earlier and more accurate diagnosis.”
Ian Abbs, Guy’s and St Thomas’ NHS Foundation Trust
Ian Abbs, CEO of Guy’s and St Thomas’ NHS Foundation Trust, said having access to the compute resources and AI capabilities of Cambridge-1 looked set to have a transformative impact on the diagnosis and treatment rates for patients.
“The power of artificial intelligence in healthcare will help to speed up diagnosis for patients, improve services such as breast cancer screening, and support the way that we risk assess and prioritise patients according to clinical need,” said Abbs.
“The Cambridge-1 datacentre…will enable us to be among the first to benefit from these new AI capabilities – using the very latest technology to benefit our patients, as well as manage precious resources more efficiently.”
Oxford Nanopore Technologies is an organisation working towards enabling the genomic analysis of any living creature, with its work spanning a wide variety of research areas, including human and plant health.
With this in mind, the organisation uses Nvidia technology in its various genomic sequencing platforms to develop AI tools that will improve the speed and accuracy of its genomic analysis, so that its algorithmic processing times take hours rather than days.
“Harnessing the power of Cambridge-1 will help us further speed up our algorithm development to support powerful, accurate genomic analysis,” said Rosemary Sinclair Dokos, vice-president of product and programme management at Oxford Nanopore.
“This will, in turn, enable the scientists using our technology on the ground to gain more insights than ever before, across a breadth of research areas.”
Looking beyond Cambridge-1
The Cambridge-1 supercomputer is located in the Harlow-based campus of life sciences-focused colocation provider Kao Data. It is the first supercomputer of its kind that Nvidia has built for use by external research teams.
“It’s very difficult for organisations in the healthcare space to process that [healthcare] data and draw clinical conclusions from it without the application of compute and something [on the scale] of Cambridge-1,” added Hogan.
But as well as being accessible to UK-based healthcare and research teams, the build-out is also designed to benefit researchers and patients across the globe, he continued.
“The whole idea of Cambridge-1 is to apply supercomputing to the healthcare ecosystem in collaboration, by working closely with partners to develop with these world-leading individuals and organisations mechanisms to bring AI to the point of care, and ultimately transform the lives of people throughout the world,” said Hogan.
“And whilst this is a UK-based system, and will be very much transformative to the UK, the benefits will be seen around the world.”
As previously reported by Computer Weekly, the installation of the Cambridge-1 supercomputer at the Kao Data site occurred just 20 weeks after the inception of the project. This represents a markedly accelerated pace of deployment, given such environments usually take several years of planning and building to bring online.
The reason why it was possible to bring Cambridge-1 online in 20 weeks is due to the modular nature of the supercomputer’s architecture, added Hogan.
David Hogan, Nvidia
Nvidia said it plans to follow up the Cambridge-1 build with the creation of an AI centre for excellence, which will be home to a supercomputer powered by technology made by chipmaker Arm.
The centre is intended to serve as a collaboration hub for AI researchers, scientists and startups across the UK, and users of it will be able to benefit from access to Cambridge-1, as well as the compute capabilities of other supercomputers across the UK as they come online too.
In the meantime, Hogan said during the pre-brief that it is Nvidia’s hope that Cambridge-1 will become “the blueprint” for how to deliver AI projects at scale.
“There really isn’t a platform out there that can really bring in an industrial approach to how you address healthcare, and it’s going to place the UK in a world-leading position to be able to demonstrate how you go on to adopt this type of technology and apply it to the healthcare ecosystem. I’m sure we’ll see many centres around the world look to adopt a similar approach,” he added.