Hey presto! the quantum supercomputer

Mythical cats in boxes? Time travel? How about computers as powerful as the Universe? We're talking quantum physics. And now the technology is with us
Click to follow
The Independent Online

Quantum physics used to be exotic science. It is now technology. That doesn't mean that we understand it - as Richard Feynman, one of the greatest quantum mechanics, was fond of saying, "nobody understands quantum physics". Don't try to understand how the quantum world can be the way it is, he urged; just lie back and enjoy it.

Quantum physics used to be exotic science. It is now technology. That doesn't mean that we understand it - as Richard Feynman, one of the greatest quantum mechanics, was fond of saying, "nobody understands quantum physics". Don't try to understand how the quantum world can be the way it is, he urged; just lie back and enjoy it.

It means that engineers have learnt how to manipulate quantum entities, such as single atoms, using the same kind of empirical, suck-it-and-see approach that enabled engineers to design effective steam engines long before the laws of thermodynamics were worked out. One of those engineers, Terry Clark, of the University of Sussex, believes that as a result we stand on the threshold of a new age of technological advance, where Arthur C Clarke's adage "any sufficiently advanced technology is indistinguishable from magic" will come true before our very eyes.

It isn't just Terry Clark who thinks the quantum revolution is at hand. His team has just received major recognition in the form of funding from the new National Endowment for Science, Technology and the Arts - one of just eight such awards made in the first year of its existence. The aim of the team is to develop machines operating according to quantum mechanics rather than Newtonian mechanics.

"This could lead to dramatic revolutions in technology in the next century, at least on a par with those we have experienced this century," says Clark. To give some idea just how dramatic, this includes the possibility of teleportation (not quite Star Trek style, but transporting particles from one place to another instantaneously). That, however, may lie a decade or two down the line.

Right now, though, one of the key areas in which this revolution is developing is in the field of computers, with the prospect of quantum computers millions of times more powerful than any machine available today. But anyone familiar with computers may be startled to learn that one reason for all the excitement, according to the Oxford-based quantum physicist David Deutsch, is that we already have the technology to build a four-bit quantum computer. In case you need reminding, four bits - four binary digits - correspond to just half a byte, in the familiar terminology.

The excitement arises because the computer in question operates on quantum-mechanical principles, not in line with the physics of everyday common sense. Which makes a quantum computer, in principle, almost unimaginably more powerful than its conventional counterpart.

The best way to get a handle on this is to look closely at those individual bits. A computer bit is just an on-off switch that can exist in either of two states, so that a string of bits can be represented as a string of zeros and ones in binary code. But the "switches" used in quantum computers behave in a significantly different way. They are called "qbits", and they behave rather like Schrödinger's famous cat. This is where Feynman's exhortation to "lie back and enjoy it" comes in. What you are about to read may be weird, but it really is the way the quantum world works - and nobody, not even Feynman, has ever understood how it can be like that, so you are in good company.

Erwin Schrödinger, one of the quantum pioneers, came up with his theoretical puzzle in the Thirties, to demonstrate what he saw as the unacceptable face of quantum physics. The eponymous feline is imagined to be locked in a box with a quantum device that has a 50:50 chance of triggering the cat's death. According to all the equations of quantum mechanics (including, to his intense disgust, the ones discovered by Schrödinger himself), quantum entities such as atoms do not "decide" what state they are in until they interact with something, or somebody measures them. By imagining a quantum trigger to decide the fate of the cat, Schrödinger scaled this quantum indecisiveness up to a human (or feline) level. The traditional interpretation of quantum theory says that the cat itself therefore exists in a "superposition of states", neither alive nor dead, until somebody looks inside the box. A variation on the theme says that when faced with such a quantum choice, the entire Universe splits into two realities, one with a dead cat and one with a live cat.

Now apply that to quantum bits. Instead of being "on" or "off", a qbit (which might be an atom in a certain pair of states, or a photon) can exist in a mixture of both states. Or, if you prefer the rival interpretation, if faced with a choice of being on or off, a qbit divides (dividing the entire Universe with it) into two copies, one "on" and one "off". If you have a string of qbits (even a string as short as four), this means that in some sense they exist in every possible "on-off" combination at once - as if they are carrying out every possible calculation simultaneously.

Physicists still argue about what this means - whether it is just some fuzzy quantum superposition (easier to believe for atoms than for cats!), or whether it means there is an entire Universe for every possible combination of "on" and "off" switching in the string of qbits. Or maybe both "explanations" are wrong. Remember, "nobody understands quantum physics". But there is no doubt about the practical implications. Whereas with a conventional computer, if you double the number of bits you essentially double the power of the computer, if you double the number of qbits in a quantum computer you square its power, tripling the number of qbits cubes its power, and so on. So a machine with six qbits working together isn't three times as powerful as a machine with two qbits, but eight times as powerful.

It may not sound all that dramatic. But exponentials quickly run away with you. So, crudely speaking, a quantum computer containing just 10 qbits would be as powerful as a conventional machine containing 1,000 bits (a kilobit, one-eighth of a kilobyte). But the most amazing thing is that if you feed a problem into such a quantum computer, it behaves as if all the calculations are carried out on this "supercomputer", then gives you the answer on your simple little machine.

Where does all this computing power come from? It looks as if you have got something for nothing - or, at least, 1,000 bits for the cost of 10. But David Deutsch, a leading proponent of the "many worlds" interpretation, points out that there are many copies of the computer being built by many copies of you in adjacent universes. In this case, if 100 alternative "yous" build identical 10-qbit computers, each reality gets access to a thousand-bit computer.

That's still some way short of the power of the machine we use daily, and it is tricky trying to make machines using entities such as individual atoms and photons as qbits. But Deutsch says that there are already plans for a quantum computer using 32 qbits (4 qbytes); and Paul Davies, of the University of Adelaide, calculates that a 300-qbit quantum computer would (thanks to the power of exponentials) be more powerful than what he calls "a conventional clunk clunk" computer that used every atom in the Universe as its on-off switches.

This raises an interesting point about the nature of quantum reality, which Deutsch is quick to point out. If we ever could build a computer that behaved as if it contained more atoms than there are in the entire Universe, how else could you explain what is going on except by saying that there are other, parallel universes out there, which are somehow connected to our own by quantum phenomena? To Deutsch, a theorist par excellence, the possibility of showing the many worlds idea to be correct is as good a reason as any to build the ultimate quantum computer.

Given the pace of progress with conventional computers, who is to say that all this will not happen within a human lifetime? But it won't happen using tiny things such as atoms as the qbits, just as modern computers no longer use the glass valves on which the first machines were based. Terry Clark and his colleagues have pioneered the development of devices that are big enough to see, and hold in your hand, but which behave like single quantum entities.

These superconducting quantum interference devices (or Squids) are each about the size of a wedding-ring, with a constriction at one point in the ring. When cooled to very low temperatures, and tweaked in the right way, an array of 300 of these things, working together as qbits, could well form the basis of the ultimate quantum computer. And Clark, unlike Deutsch, has his feet firmly on the ground when it comes to justifying such research: "because," he says, "new technology creates wealth". That does not necessarily mean personal wealth, if you work at a British university, but economic growth.

Clark emphasises that his group is not the only one with radical new ideas about how to make quantum machines, and says that together all the researchers working in this field have now "laid the basis for a whole new technology, that will clearly involve quantum computers". And once you've got those super-powerful quantum computers, you can use them to push the technology further - certainly to help design a working teleportation system for inanimate objects, and possibly (according to a few theorists) opening up the prospect of some form of time travel. The science exists, and those four-bit computers represent the beginnings of the new technology. Anyone old enough to remember the Sinclair ZX-81 computer may have the merest inkling of what those four bits may be worth by the year 2020.

The writer's series 'Quantum' is to be broadcast on Radio 4, beginning at 9pm on Wednesday, 1 December