This article is more than 1 year old
'Large-scale computing' needs a government team driving it, says UK.gov
Report from the Government Office for Science lays out the hurdles to supercomputing relevance
A report from the Government Office for Science has proposed the formation of an in-house team dedicated to large-scale computing, as it bemoans the nation's weak standing in the international supercomputing sector and makes a series of recommendations for improving matters.
"This report sets out the building blocks required to create a world-class computing ecosystem," claimed Sir Patrick Vallance, the government's chief scientific advisor, "so the UK can realise the true potential of large-scale computing and harness this technology to strengthen its position as a science superpower."
"The core of this report is based around seven challenge areas and options to overcome each," Vallance continued in the report's foreword. "The need for better national coordination; nurturing a healthy UK ecosystem; ensuring we have the right mix of hardware, skills, and software; minimising energy consumption; and the UK supply chain."
There may be those who view the report's key recommendation with a suspicious eye, however: "To establish a team within government to take policy responsibility for large-scale computing." Such a move could give the government increased control over the future development of large-scale computing projects in both the public and private sectors.
"A large number of government departments and arm's-length bodies have some responsibility for or involvement with large-scale computing," the report explained. "No single team within government has overall responsibility for coordinating policy on large-scale computing."
"The UK's approach to computing is uncoordinated, introducing inefficiencies into procurement and limiting sharing of resources," the report continued. "This [proposed] team would have ownership of policy for large-scale computing [and] would be responsible for developing a rolling long-term roadmap for the large-scale computing ecosystem."
Could a dedicated government team really push the industry forward?
"Yes," Simon Cox, professor of computational methods at the University of Southampton and no stranger to building large-scale computing systems for academic research, told The Register. "The reason for that is if one thinks of the infrastructure where government has understanding – take today's newspapers and the gas industry, take today's water industries, road, rail infrastructure – it will be useful, setting it up and putting [large-scale computing] on that platform as part of the infrastructure necessary to do research."
As for the make-up of the proposed specialist team, the report's authors suggested it should comprise both civil servants and secondees from the public sector, private industry, and academia, leaning on experts and providing "a forum for major users to share ideas and keep up with international trends", while also "providing policy guidance" to the government.
The team would also be given the task of addressing the seven challenge areas highlighted in the report – including the reliance on foreign technology.
"It is likely not feasible for the UK to be able to compete in all areas of the supply chain," the report admitted. "For example, entering into the field of semiconductor fabrication would require a multi-billion-pound investment in a semiconductor foundry, which would be difficult to justify.
"However, the UK has the potential to become a leader in certain areas of the market. One area is the development and prototyping of novel compound semiconductor technologies. Other areas where the UK is well-placed to capitalise on future commercialisation opportunities include hardware acceleration and novel computing paradigms (such as neuromorphic computing and quantum computing)."
- AWS taps up Singapore scientists to overcome hurdles facing quantum computing
- Useful quantum computers will be impossible without error correction. Good thing these folks are working on it
- IBM pulls up the ladder behind some supercomputer customers
- AMD opens wallet to lure scientific computing boffins away from Nvidia's CUDA onto its Instinct accelerators
Despite admitting that Britain's large-scale computing ecosystem lags those of other nations – with just 12 systems in the current TOP500 list (including the nation's fastest system as built and maintained by US tech giant Nvidia), with those 12 systems representing just 1.4 per cent of total performance capacity – the report declared a "strong case for continued public investment", but only if "coupled with improved long-term strategic planning."
The report also weighed up the pros and cons of cloud computing, describing it as an "increasing popular access model" which particularly appeals to new users and smaller enterprises – but warned that it was a market "currently dominated by a limited number of leading providers based in the US."
One of the biggest issues highlighted in the report, however, has little to do with the underlying technology itself. "There is an acute shortage of large-scale computing professionals, including system architects, data engineers, system operations professionals and software engineers, [in the UK]," the report claimed.
"Exploitation of advanced computing requires skilled cross-disciplinary teams. Within government and academia, career paths for computing professionals are often not well-defined," the report continued, recommending improved career pathways including better job security, salary structures and progression opportunities for existing staff, while opening the field up more broadly with "non-degree entry routes and retraining opportunities [such as] apprenticeships, trainee schemes, conversion courses, and continuing education."
"Is there a skills shortage? Yes, yes, yes, and yes. It is very hard to recruit and retain staff," Cox told us, emphatically agreeing with the report's thrust. He added that, "even speaking as a professor at a Russell Group University," non-degree routes could be a potential answer.
"The right message is there," Cox concluded. "I think for the first time it reflects the true breadth and depth across the UK. If you'd read this sort of report two or three years ago, you perhaps wouldn't see data quite so prominently mentioned. Data is part of the story.
"I gave a talk to a prize day, and I used that old phrase 'data is the new oil,' which is so hackneyed and old now that you almost think nothing of it. But you can't talk about high-performance computing without adding data, and I did see more of that written through the report than perhaps previously.
"Next time around will data be even more on people's minds? That's perhaps me being an academic, thinking 'what would you do for the next version of such a report', that's by no means a critique of what's been written here which I think is a jolly strong document. I'm pleased to see where they'd picked up on some things that have been on my mind for a few years."
The report, as is made clear in a header on every single page, "is not a statement of government policy," and thus far the government has not indicated which – if any – of its recommendations it intends to take on board.
The full report is available to download now from GOV.UK. ®