Friday, February 15, 2013

Big data and energy efficient computing: where will the money go?

At the end of last month, in a speech at the UK Policy Exchange, Minister for Universities and Science, David Willetts, described how the ?600m provision for science, contained in the government?s Autumn Statement, will be used to boost growth in ?eight great technologies.? The lion?s share, ?189m, will go towards big data and energy efficient computing. So why, among the possibilites - robotics, agri-science and energy among them - has this pair of technologies been chosen, and just where will the money go?

?

Strengths

In a report, published alongside the Chancellor George Osborne?s speech to the Royal Society, three main criteria were established for a technology to be on the list: It has to an important area of scientific advance; Britain has to have a distinctive capability and it should have reached the stage where identifiable commercial opportunities are emerging.

Big data, and its necessary companion, energy efficient computing, were identified, in extensive government research, as fulfilling all these criteria. The much anticipated data deluge is now upon us, with many of the consequences and advantages already chronicled. In his report, Willetts states that:?

?Capturing value from all this data ? for economic growth and social benefits such as improved health ? requires a transformation in data analysis. With the right investments, the UK is well placed for the big data revolution.??

The skeptical will point to the fact that the UK has only 25 of the world?s 500 most powerful computers (out of 107 in Europe compared with 253 in the US, and 68 in China), however, Mr Willetts points out that crude computing power is not the most important factor here and the UK has some very distinctive strengths:

First, the UK is good at the algorithms needed to handle diverse large data sets, with historical advantages in both mathematical and computer sciences, as well as cryptography. British scientists now play a key role in research projects which generate very large data-sets, such as the search for the Higgs boson at CERN. These capabilities can be used across disciplines, so, for example, a computer programme written to analyse astronomical data is now being used by neuroscientists studying brain cells.

Secondly, the UK possesses some of the world?s best and most complete data-sets in healthcare, demographics, agriculture and the environment. Their value is increasing, with new ways to access and analyse them and the UK has become adept at statistical techniques, partly because of the amount of material generated historically.?

Lastly, harnessing and storing huge amount of data requires novel technologies and skills in energy efficient computing, with IT becoming an increasingly heavy user of energy. Efficiency requires high quality software and smart algorithms and, although the UK doesn?t have the world?s most powerful computer, it does possess some of the world?s most energy efficient super computers.?

?

Boost

To this end, part of the ?189m provision will be used to boost this capability, and ?30m has recently been earmarked to develop supercomputer software that can handle huge datasets, such as the output of ultra-sensitive radio telescopes.

This will mean that the UK is able to lead projects like handling the vast amounts of data produced by the world?s biggest radio telescope, the Square Kilometre Array (SKA), which requires new breakthroughs in computing systems and data analysis.

?11m has been dedicated to developing software for the telescope, which is based in southern Africa and Australia, and will be fully operational by 2024.?

Power-efficient computing technologies and software will be capable of handling the extensive datasets created by large experimental research initiatives. A portion of the money (?19m) will go to the Science and Technology Facilities Council?s (STFC?s) Hartree Centre, at Daresbury, to support its work on power-efficient computing technologies. STFC chief executive Prof John Womersley said:?

?For industry, this could mean extreme modelling for smart materials for industrial adhesives or coatings, or in the engineering and manufacturing for the car and aerospace industries. For the consumer, it could result in longer-life mobile communications for phones and tablet computers".

Researchers at Hartree, home of the UK?s most powerful working supercomputer, the IBM BlueGene/Q, are already working with IBM and Intel to develop new technology. Their tools have also helped commercial partners, such as Bentley, which has used the centre?s 3D Visualisation Suite, to reduce the number of prototypes needed in the development of a new car.

New research will focus on three areas: The centre will aim to develop hardware that can achieve more flops per Watt from computer systems; it will also try to understand the performance of algorithms and software by profiling performance and energy consumption. Finally, the centre will look at ways to lower the power usage effectiveness (PUE) of machine rooms by studying different architectures.?

According to Prof Womersley: ?This investment will enable the development of new, more capable and more energy-efficient computing for an immense range of applications".

This means benefits for many sectors such as social science studies and the Economic and Social Research Council led Life Study, the most ambitious birth cohort study yet, which will track around 100,000 children from birth. Arts and humanities will also benefit, as, for example, it becomes possible to have an accessible record of every work of art in a British public collection.

The aim is to establish the UK as a world leader in energy-efficient supercomputer software and also to provide the tools to help industry carry out advanced modelling techniques to test new materials or aircraft designs.

"Businesses will invest more as they see us invest more in computational infrastructure to capture and analyse data flows released by the open data revolution," said Mr Willetts, but he warned:?

"Some of the technologies for which we have high hopes today will turn out to be clunkers tomorrow. That is because this is all about taking risk -- if the risk was much lower then we could indeed leave it to straightforward business decisions."

Mr Willetts' full speech, including the full list of eight technologies the government is investing in, is available here.

Current developments and the future R&D needs of SKA are the basis of an workshop which takes place on 20 March 2013 at the Jodrell Bank Observatory in Cheshire. More info available here.?

?

?

Source: https://connect.innovateuk.org/web/ictktn/articles/-/blogs/10868121

dodgers triple play baa samoyed kenny powers kenny powers carl hagelin triple play

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.