UK researchers primed to build next-generation supercomputer

By Gareth Morgan

21 Sep 2011

Comments: 3

High-performance computer

The UK’s Science and Technology Facilities Council (STFC) is set to invest up to £16m to create a new supercomputer, based on the principles of so-called exascale computing.

The STFC, one of seven UK research councils, is charged with ensuring that the UK can play a leading role in scientific research, with a specific focus on areas such as atomic and molecular physics, computational chemistry and bioscience.

Further reading

It already works with experts in high-performance computing (HPC) in developing software capable of running on petascale supercomputers. But it wants its new system to be built on technologies that can reach exascale levels.

To put those performance levels into perspective, the latest measurements of the world’s fastest computers put Japan’s K Computer top. It achieved 8.2 quadrillion floating-point operations per second – or 8.2 petaflops.

The new STFC system, which will consist of a minimum of 2,048 x86 cores, will not reach anything like such speeds. But STFC hopes that the work undertaken on the system will help advance the use of HPC within academic research.

The contract for building the system is expected to be finalised by March 2012.

Reader comments

Where is the next generation?

2048 x86 cores means around 170 conventional 12-core nodes. At a cost of around £5K per node you get £850K. Now, how do we spend the remaining £15 millions? Wouldn't be wise to help research and innovation other than investing in a technology that has big problems of scalability?

Posted by: diego  21 Dec 2011

Exascale computing is the natural next step

Supercomputers have come a considerable distance since they first appeared in the 1960s. Exascale computing is the natural next step in this progression and offers incredibly exciting opportunities for the STFC and the wider scientific community in the UK. For a relatively small island we’re already crunching a tremendous amount of scientific data and running immeasurable calculation-intensive tasks every single day. Advance those computing capabilities beyond the petascale and things get really interesting! Who knows, in a few years time we might be looking back on the STFC supercomputer as the machine that discovered the cure for cancer, found a vaccination for the AIDS virus or unlocked the potential of fusion power.

Posted by: Christopher Nathan, Platform Computing  27 Sep 2011

Need Help? Just ask.

If the designers are having trouble with cooling the system, tell them they only have to ask.

I solved the problem 12 years ago.

http://www.linkedin.com/groupItem?view=&gid=87791&type=member&item=70970324&qid=452bd879-ab9e-413d-ae14-70f655845708&trk=group_most_popular-0-b-ttl&goback=.gmp_87791

Sincerely,

Paul Mead

Posted by: Paul Mead  22 Sep 2011

Have your say on this article

All fields required. Your email address will not be displayed on the site.

By submitting a comment you agree to abide by our Terms & Conditions

  • Digg
  • Tweet

Newsletters

Sign up for our FREE newsletters

Technology Patent Wars

Large companies such as Microsoft, Facebook and Google have been hoovering up technology patents recently. Is this stifling innovation?

87 %

5 %

8 %