For some time, industry experts have been warning that Moore’s Law, that’s the phrase that is used to sum up the way in which computers get faster and more powerful over time, is coming to an end. If that is so, the implications are far-reaching. But no so fast, researchers beavering away at the University of Central Florida reckon they may have come up with a way to make computers 1,000 times faster – that’s 1,000 times, think of the implications of that!
The key to the breakthrough is graphene, the one-atom-thick carbon-based super material, first isolated at the University of Manchester in 2004.
Before the explanation begins, let step back and recall the story of Moore’s Law. It is named after Gordon Moore – co-founder of Intel, who in 1965 said that the number of transistors on an integrated circuit were set to double ever two years. At first, he only thought that this pattern would last ten years, but then, over time, Moore’s Law was tweaked to become short-hand for saying computers double in speed every 18 months.
And we underestimated the implication of this. Thanks to Moore’s Law, a single, bog standard, smartphone has an order of magnitude more processing power that was available to NASA during the moon landings.
Maybe, though, we forget too easily how surprising the consequences of this have been. People laughed at Bill Gates’ dream of a PC on every desk, but we have since moved onto a computer in every pocket and now a computer in billions of sensors – each packing more punch than a supercomputer a decade or so ago.
But Moore’s Law depends on size – or rather on things getting smaller. The more you can squeeze in on an integrated circuit the faster and more powerful the resulting chip. But silicon has its limits – it seems we are approaching the point when a silicon-based chip just can’t get much faster.
The implications are profound – will we still want to upgrade our smartphones or PCs every few years if they are not getting any increase in power?
The possible solutions are manyfold. Specialist chips, designed for specific tasks may beef up computers. We are likely to make greater use of the cloud – with all the heavy processing, much of it carried out by specialist chips, carried out on a server, or servers somewhere. And then there is quantum computing – this is exciting, and slightly scary technology, but is probably many years away and in any case, will be suitable for all tasks.
In theory, graphene is another potential answer. It’s a super-efficient conductor – too efficient some say, unlike silicon which is a semiconductor – you can’t turn a graphene transistor on and off like you can with silicon.
But the team of researchers at Florida have worked out how to use graphene-based transistors by applying a magnetic field to a graphene ribbon. “Increasing or decreasing the strength of the magnetic field would also increase or decrease the flow of current through this new kind of transistor, much like a valve controlling the flow of water through a pipe,” states Science Daily.
To cut to the chase the implications are as follows:
The speed of computers applying this technique could be 1,000 times faster.
But the technology is theoretical, it may be several years, even more than a decade before it works in the real world.
But pause and consider the implications – whether it happens in five, ten or twenty years, what kind of things could a computer that is 1,000 times faster that the current crop be able to do? The mind boggles.