Log in
This case study describes the development, application and commercialisation of an open source tool, BSMBench that enables supercomputer vendors and computing centres to benchmark their system's performance. It comprehensively informs the design and testing of new computing architectures well beyond other benchmarking tools on the market, such as Linpack.
The significance of our code is that, unlike other benchmarking tools, it interpolates from a communication- to a computation-dominated regime simply by varying the (physics) parameters in the code, thus providing a perfect benchmark suite to test the response of modern multi-CPU systems along this axis. The impact of this work has great reach: a start-up company, BSMbench Ltd, has been founded to develop and commercialise the software; adopters have included IBM - one of the giants of the supercomputer world (where it uncovered errors in their compilers); it has been deployed by Fujitsu to validate its systems, by HPC Wales, a multi-site, commercially focussed national computer centre and by Transtec, an HPC company employing over 150 staff; and tutorial articles about BSMBench have appeared in magazines such as Linux Format.
This software tool spawned from our research into "Beyond the Standard Model" (BSM) physics which aims to understand the Higgs mechanism in particle physics at a fundamental level. This involved simulating quantum field theories using bespoke code on some of the fastest supercomputers on the planet.
The Abraham solvation parameter approach developed at UCL has become integral to the work carried out by drug discovery teams at [text removed for publication] and other major pharmaceutical companies, as well as research and development groups at international chemical companies including Syngenta and [text removed for publication]. It enables chemists to predict physicochemical and biochemical properties of chemicals, including drugs and agrochemicals, rapidly and efficiently, without the need to conduct time-consuming experiments. The method helps drug discovery teams to identify and optimise the most promising compounds, and often results in fewer compounds being made before a candidate is selected, saving time and resources. The approach has been integrated into software used for drug discovery [text removed for publication].
In partnership with the US company Nalco, the University's Surfactant & Colloid Group developed a new multifunctional technology (Clean n Cor) for the oil industry that both removes accumulated deposits at a metal surface (enabling "break-through" of corrosion inhibitor to the metal surface) and inhibits corrosion. Clean n Cor technology not only protects assets such as oil pipelines against corrosion but also maximises oil production through enhancing water injectivity (water flow per unit pressure drop). Since its launch in 2007, it is currently one of Nalco's fastest growing new technologies and is used at over 100 production locations worldwide.
The performance of absolute distance measuring systems has been improved in terms of accuracy, traceability, reliability and cost through the introduction of new methodology arising from research at the University of Oxford. This has brought commercial benefit to a German company making measurement systems, through the creation of a new product line. New capabilities for measurement have been delivered to a first customer in Germany. The research has also resulted in the establishment of new activity at the National Physical Laboratory, and influenced UK and European technology roadmaps for future manufacturing.
Research at Lancaster led to a novel approach to detect the source of cases of campylobacteriosis (a bacterial foodborne disease). The application of this method to data from New Zealand pin- pointed that New-Zealand's high rate of cases was linked to the eating of contaminated poultry. These results were a key part of the evidence used by New Zealand's Food Safety Authority to introduce a new code of practice for the poultry industry. The impact of this code of practice has been a halving of the number of reported cases of campylobacteriosis in New Zealand (from around 16,000 cases in 2006 to less than 7,000 in 2008). With notification rates estimated as 1 in 10, this corresponds to around 90,000 fewer actual cases per year. The saving for the New Zealand economy during the REF census period has been independently estimated as between £100M and £150M.
Our research on Active Shape Models (ASMs) and Active Appearance Models (AAMs) opened up a radically new approach to automated image interpretation, with applications in industrial inspection, medical image analysis, and face tracking/recognition. We identify:
Series of images from animal behaviour studies contain vast quantities of complex and highly valuable data. Extracting the value from this scientific data often requires expert annotation. This is frequently an intuitive process based on experience gained through years of training to make important decisions. Experts are rare, expensive and hard to train so the iBehave project at the University of Edinburgh (2006-2009) sought to reverse this model and deliver systems that learned to mimic expert annotation of video data. This effort resulted in a new spinout from the School of Informatics, a software company called Actual Analytics Ltd (Actual). Founded in 2010, Actual delivers innovative software solutions for behaviour analytics which use machine learning algorithms to process video data of laboratory animals to improve the accuracy of the experimental process and reduce the need to use animals in scientific research.
Nearly every large-vocabulary speech recognition system in current use employs outputs from fundamental research carried out in the University of Cambridge Department of Engineering (DoEng) on adaptation of Hidden Markov Models (HMMs). One example of the commercial application of these outputs is their use on the Microsoft Windows desktop for both the command and control functions and the dictation functions. Approximately one billion copies of Windows have been shipped since 2008. Other examples show the outputs used in the automatic transcription of a wide range of types of data. [text removed for publication]
The largest investment banks in London each have thousands of servers largely devoted to Monte Carlo simulations, and to quantify their risks and satisfy regulatory demands they need to be able to calculate huge numbers of sensitivities (defined below) known collectively as "Greeks". An adjoint technique developed by Professor Mike Giles in 2006 greatly reduced the computational complexity of these calculations. The technique is used extensively by Credit Suisse and other major banks, reducing their computing costs and energy consumption. It has also led to the Numerical Algorithms Group developing new software to support the banks in exploiting this new adjoint approach to computing sensitivities.
Using advanced mathematics and numerical modelling we have demonstrated how fundamental understanding of laminar-turbulent transitions in fluid flows can save energy. From 2008 we helped the cleantech company, Maxsys Fuel Systems Ltd, to understand and improve their technology and demonstrate to customers how it can reduce fuel use by 5-8%. Customers including Ford Motor, Dow Chemical and Findus testify to the impact from financial savings and reduced carbon emissions obtained by installing Maxsys products on industrial burners used widely in many industrial sectors including automotive, bulk chemicals and food. In 2010, Selas Heat Technology Company bought the Maxsys brand to invest in this success.