The UK claims a host of skills in computer science, mathematics and numerical algorithm development, and British scientists and engineers are making significant contributions to some of the largest and most complex experiments in the world.
These experiments more often than not rely on high-performance computing (HPC) to manage and process the resulting data, such as the Large Hadron Collider at Cern, and the world’s biggest radio telescope – the Square Kilometre Array – which will generate enough raw data every day to fill 15 million 64GB iPods.
As supercomputing becomes more accessible, it is more important than ever to ensure computing skills are given the support required to continue and expand on this work. Businesses, researchers, academics and students need these new skills plus access to the latest technology available in the field of supercomputing – such as high-performance number-crunching servers or graphically intense professional software – in order to support advances in science and drive improvements to products and services, which in turn contribute to the wider economy.
Supercomputing is no longer the reserve of a few select universities, but rather on the priority list for both the government and education bodies. In January 2013 David Willetts, the minister for universities and science, published a document called Eight Great Technologies that supported the allocation of £600m of additional funding for science. £156m of this went to UK research and development, of which a significant proportion has been invested in new academic supercomputers, giving academics and researchers the power of HPC needed to carry out vital research in highly efficient timescales.
Making use of HPC improves the quality of their output, while reducing time to value, therefore making a more productive use of their research and development spend. Training and support for students in the use of parallel computing techniques must be provided to ensure that HPC is being used effectively to ensure research data is to be collected, analysed, processed and understood correctly.
Nvidia runs the established Compute Unified Device Architecture (CUDA) Centres of Excellence (CCOE) program, which recognises, rewards and fosters collaboration with institutions at the forefront of massively parallel many-core computing research.
Currently available at the Universities of Oxford and Cambridge, the centres demonstrate a unique vision for improving the technology and application of parallel computing by developing the state of parallel computing education, in order to train the next generation of computer scientists and computational scientists for a world of massively parallel computing.
The aim is for universities and Nvidia to work collaboratively to decide on which resources suit each institution’s educational offering. This can be achieved in a number of ways, from providing educational materials and equipment support – including pre-release hardware and software, which allows research on next-generation technology – to training workshops which teach students to be proficient in CUDA C/C++, Nvidia’s parallel computing platform and programming model.
The initiative also helps to establish research, educational, and recruiting relationships with the foremost academic institutions in the world. Having already supported a number of students in the UK to date, these centres can support the education of the next-generation workforce all the way through their careers, empowering academics to do world-changing research, by dramatically increasing the computing power available to scientists and engineers – on the desktop, in the laboratory, and in the data centre.
UK organisations are using HPC to make significant advances in engineering and manufacturing, environmental studies, financial services, life sciences and materials, as well as fundamental scientific research. Willetts acknowledged that the UK must “out-compute to out-compete”.
Going forward, offering substantial support and incentives will be intrinsic to enticing more people into the supercomputing sector, and the more cohesive businesses can be, the greater to benefit to all parties involved.
This can be as simple as recruiting other technology providers to work alongside existing schemes, which is currently the case with CCOE, as these institutions are always seeking partnerships with hardware vendors to host machines for teaching laboratories and research clusters.
While there is not a lack of available talent, organisations and educational bodies need to increase investment and provide targeted training in order to take a step in the right direction towards bridging the formidable skills gap.
Sumit Gupta is general manager for Tesla GPU Accelerated Computing at Nvidia. He is writing for V3 as part of our Make IT Better campaign to improve computing learning in schools.