The supercomputer arms race has just taken a big step up, in political terms, after US president Barack Obama penned a plan to build the world’s first exascale computer.
To explain an exascale computer – which can run at 1,000 ‘petaflops’ – we should probably give context.
Currently, the world’s fastest supercomputer sits in China, with the Tianhe-2 running at 33.86 petaflops, meaning it can perform 33.86 quadrillion arithmetic operations per second.
The US, currently, has the second-fastest machine, which runs at about half that.
True, Intel is currently working on the US’ upgrade on that, a 180-petaflop machine due in three years. Called Aurora, the machine will be used for scientific research when it is released in 2018.
However, Obama’s call for something nearly six times the size of that as-yet-unfinished US$200m project shows a serious sign of intent.
Many interested parties
Establishing a new National Strategic Computing Initiative (NCSI), Obama is drawing in a swathe of agencies to liaise on a project to accelerate the delivery of a “capable exascale computing system” aimed “approximately 100 times the performance of current 10 petaflop systems”.
The US Department of Energy, Department of Defence and National Science Foundations will lead a project that Obama phrases quite tellingly.
The NCSI will establish, over the next 15 years, the future of super-computing “even after the limits of current semiconductor technology are reached”, which is essentially the post-Moore’s Law era.
Over the next three months the first plan will be formatted, with yearly reports to follow.
There is no timeframe for the project and it’s not all federally focused, with a call to include the private sector with its creation, completion and operation, too.
A growing trend
Late last year the US Intelligence Advanced Research Projects Activity (IARPA) announced a multi-year research effort to develop a superconducting computer.
If successful, technology developed under the cryogenic computer complexity (C3) programme will pave the way for a new generation of superconducting supercomputers that are far more energy efficient.
“Computers based on superconducting logic integrated with new kinds of cryogenic memory will allow expansion of current computing facilities while staying within space and energy budgets, and may enable supercomputer development beyond the exascale,” said Marc Manheimer, C3 programme manager at IARPA.
The US currently houses 233 of the world’s 500 fastest computers, Europe has 141 and China has 37.
Main image via Shutterstock