What can take the place of Moore’s Law once the limit is reached?
This five year Superspin project is funded by the Engineering and Physical Sciences Research Council
The chips are nearly down for Moore’s law but what will take its place? The laws of physics decree that performance gains cannot be maintained by shrinking processes for ever and quantum architectures still lack maturity.
So, we’re currently skirting the limits of Moore’s 40-year-old law. But, thanks to new spintronics technology coming out of the UK’s Cambridge University, tantalising new architecture opportunities are opening up, including the possibility of ultra low power data centres.
Spintronics is not new. In a nutshell, it’s a method of spinning electronics in various ways to alter their magnetic and electrical properties. What is new is the combination of spintronics with ultra energy-efficient superconductors, a pairing previously believed to be impossible. But that’s exactly what the Cambridge team have managed to achieve, thus enabling an entirely new generation of processing technology.
The Cambridge research, which goes by the name of ‘Superspin’ is focussed on developing devices that eliminate energy loss without driving up cooling requirements to the point where the net gain in performance and energy efficiency is lost. The research is still in a very nascent stage, with a working prototype at least five years away, but it has potential to dramatically cut the energy costs of high-performance computing.
Spintronics stores and processes data as the spin of electrons, and could, in principle, save power compared with electronics because it avoids moving charge around. It is also potentially fast. Very fast.
The aim of the £2.7m five-year project, funded by the Engineering and Physical Sciences Research Council (EPSRC), is to use this as the basis for a computing architecture.
Initial stages will explore different ways in which spin can be transported and magnetism controlled in a superconducting state.
“Although work is already underway in several other countries to exploit superconducting spintronics, the Superspin project is unprecedented in terms of its magnitude and scope,” said the University. “Researchers will explore how the technology could be applied in future computing as a whole, examining fundamental problems such as spin generation and flow, and data storage, while also developing sample devices.”
Data centres are the engines of the digital economy. But they are also very energy intensive with estimates that some three per cent of the power generated in Europe is already being used by data centres. “Our basic calculations suggest that superconducting spintronics will be massively more energy efficient than current spintronics,” the researchers say.
However, an equally interesting driver for the research is to use superspin as a possible alternative to semiconductor technology — as a new route to sustain Moore’s Law of shrinking electronics. Spintronics proposes utilising the spin alignment of electrons as a medium to store (the 0 or 1 of) digital data.
“Information technology now is based on such small objects you just can’t use conventional superconductors,” notes Dr Jason Robinson, one of the project leads. “By combining superconductivity with spintronics it’s not only that you can create circuits without [energy] dissipation, but it’s that you create new physics. So that means there’s a lot of new opportunities created through this combination. There’s a lot of undiscovered physics to be explored. It introduces lots of new ideas that were not possible previously. So that’s exciting, and indeed a large part of our grant is to develop the science of those other areas as well.”
By 2021, the team hopes to have made sample logic and memory devices.
To allow industry to make use of discoveries, the project will report to an advisory board with representatives from several technology firms, said the University.
“It’s important to understand that this is the first ever superconductivity and spintronics funded program,” adds Robinson. “The way the grant has been set up in the first three years, there’s a series of parallel projects. Some are more applications biased than others but the application stuff has to develop alongside the science… Everything we do is moving us towards the prototype. It’s a fundamental program with the aim of triggering applications in spintronics.”