Supercomputers power scientific discovery, sustainable design, and next generation exabyte systems at one of the world’s oldest universities
Founded in 1209, the University of Cambridge in England is the world’s third-oldest existing university and has a long and distinguished history of scientific discovery. The many brilliant scientific minds who have taught and conducted research at Cambridge include: Charles Darwin, who formulated the theory of evolution; Francis Crick and James Watson, who mapped the structure of proteins in DNA; and Stephen Hawking, the theoretical physicist whose work on the origins and structure of the universe revolutionized the field.
Today, one of the hotbeds of scientific research and innovation is within the University of Cambridge Research Computing Services, where efforts to improve medicine, treatment, and healthcare outcomes are benefiting from simulations, artificial intelligence techniques, and data analytics on supercomputers.
Data-intensive Healthcare Research
IDC estimates that an average of 270 GB of healthcare and life science data is created for every person in the world and that 30% of the world’s data volume is generated by the healthcare industry. At the University of Cambridge, with some of the world’s most sophisticated supercomputers, more than 3,000 researchers work on over 700 projects with large volumes of patient data on 2,500 servers.
One project lets healthcare professionals analyze a patient’s genome to understand how it may influence the course and treatment of a particular disease. For example, a swift diagnosis and treatment plan can be tailored to a patient based on analysis of data on specific types of cancer. A consortium of biologists and genome scientists are working to generate a detailed gene variant analysis of 10,000 people to identify the links between gene variation and illness. Another project provides COVID insights based on national healthcare data on a weekly basis to help policymakers in the U.K. make better informed decisions on public health initiatives.
Sustainable Computing
Reducing the supercomputing center’s environmental impact is another major goal at the University. To help lower energy use, researchers are writing code that estimates the amount of time and energy an operation takes and to include power consumption as part of project estimates. Developers could be incentivized to write code that makes more efficient use of supercomputing cycles and specific higher-energy-use components. Other sustainability efforts include exploration of ways to reduce the power required for heating and cooling of hardware, including component-level cooling.
The University has nearly doubled the energy efficiency of its Wilkes3 supercomputer and it is now rated one of the most energy efficient supercomputer in the world on the International Green 500 list.
In another area of sustainability research, the University is collaborating with the UK Atomic Energy Authority (UKAEA) to design nuclear fusion reactors to try to help alleviate the climate crisis. Using supercomputing resources and big data analytics, the goal is to build carbon-neutral reactors.
Exascale Computers
Total processing power for some of the fastest supercomputers at the center are measured in petaflops ― 10 petaflops is 10 quadrillion (thousand trillion) floating point operations per second (FLOPS). These are next generation exascale supercomputers, capable of a billion billion calculations per second. They are 50 times more powerful than systems used today and combine HPC with ultra-fast data storage for analytics, artificial intelligence, and simulations.
Today, building an exascale system requires half a million servers so researchers and partner companies are building new, more powerful, computational technologies as an alternative. Large-scale, solid-state storage systems that maximize I/O performance and high performance networking solutions that provide broad interoperability are also part of exascale development efforts at the University.
Another project within the Cambridge Exascale Lab is scaling cloud-native, software-defined supercomputing with OpenStack, allowing scientists and researchers to deploy clusters on demand. The Cambridge Exascale Lab features partnerships with leading agencies and technology companies working with exascale technologies. Their systems will support some of the most computationally intensive, advanced research in the world, from understanding the universe, designing new materials and clean energy solutions, and developing personalized, data-driven healthcare.
As data volumes mushroom and artificial intelligence techniques grow more powerful, the University of Cambridge is a leader in the use of vast data sets and HPC to come up with solutions to some of the most difficult challenges facing mankind. With the digitization of healthcare in particular, University researchers are convinced that the treatment and prevention of disease and the keys to health and longevity lurk within data. Their work defines a new era in scientific discovery that may rival the work of famous Cambridge scientists over eight centuries.
For more information, read “HPC Technology Helps Scientists Tackle Tomorrow’s Problems, Today” at the University of Cambridge and/or watch the video.