We will be asking some powerful questions about the new and upcoming quantum computers!
We may investigate how quantum computers work! So stick around!!
Quantum computers promise to revolutionise the digital world, but how do you tell if a computer really is harnessing the power quantum mechanics? It’s a question that has plagued the only computer manufacturer claiming to produce quantum-powered machines – D-Wave Systems of Burnaby in British Columbia, Canada – since they went on sale.
Today, the publication of further inconclusive tests of the machines is a reminder of just how difficult it is to get an answer. We explain why it’s so hard to test a quantum computer – and whether we’ll ever get an answer to the D-Wave conundrum.
What is quantum computing and why should I care?
Quantum objects can be in multiple states at once, a property known as superposition. This means a quantum bit (qubit), the basic unit of information in computing, can be both a 0 and a 1 at the same time. Theoretically a computer with a large number of these qubits should be able to complete certain tasks, such as factoring numbers or searching large databases, much faster than their ordinary equivalent.
Has anyone built a quantum computer?
Labs around the world have built devices with a handful of working qubits, but they wouldn’t even put a pocket calculator out of business: one of the most impressive results to date is factoring 21 into 3 and 7.
Meanwhile, several years ago, D-Wave burst on to the scene, offering up its machines for sale. But despite high-profile customers – including US defence firm Lockheed Martin and Google, which operates its D-Wave machine in partnership with NASA – there are still questions about whether the machines really count as quantum computers.
They rely on an alternative theory called adiabatic quantum computing and no one knows whether the theoretical quantum speed-up this provides can be translated to real-world machines.
So is it quantum or not?
D-Wave has demonstrated that its machine behaves in a quantum way and that it can compute things, but the jury is still out on whether it is actually using quantum mechanics to hasten its calculations. “Nobody knows whether it works. It is a totally high-risk, speculative project,” says Matthias Troyer of ETH Zurich in Switzerland. “If it pays off, it is a huge breakthrough.” Earlier this year, Troyer’s team released results from tests of a D-Wave Two machine that suggested there was no quantum speed-up. Today, these results are published in Science (DOI: 10.1126/science.1252319).
Wait, why would anyone buy a computer if they don’t know that it works as claimed?
D-Wave does not publically list the cost of its computers, but they are thought to be $10 to $15 million – a drop in the bucket for a multibillion dollar company like Google. Essentially D-Wave’s customers and investors are hoping to get in on the ground floor of a computing revolution.
Can’t you just test whether it runs faster than a regular computer?
Ye-es, but first you have to figure out what kind of test to run. It has to be a fair fight – one D-Wave-sponsored test that showed apparent gains was later criticised for pitting a specialised quantum algorithm against unoptimised software on an ordinary PC.
What other aspects are necessary for a useful comparison?
The test also has to involve a problem where being quantum actually gives you an advantage. D-Wave computers solve problems in a process similar to exploring a hilly landscape where the lowest points corresponds to the best solution. While an ordinary computer is forced to climb up and over the hills to find the low points, a quantum machine can simply tunnel its way through.
The trouble is that many test problems aren’t challenging enough, leading some to suggest that the reason D-Wave didn’t show a quantum speed-up in some tests – such as Troyer’s – isn’t because it is not able to deliver a better performance, but rather because the test didn’t force it too. “The D-Wave machine would rather use classical resources instead of quantum,” suggestsVadim Smelyanskiy of NASA’s Quantum Artificial Intelligence Laboratory in Mountain View, California, which hosts the Google-purchased computer. D-Wave claims that is the case with Troyer’s test. “Those problems are simply too easy,” says Colin Williams of D-Wave.
Will this one ever be resolved?
Smelyanskiy is currently working with others at NASA and Google to develop tests he hopes will put the machine through its paces, which he presented at the Adiabatic Quantum Computing conference in Los Angeles last week. “You want to construct those tall mountains and absolutely be sure that there is no way around,” he says. “For those problems, we will be able to see if the machine really is forced to do something quantum.”
What happens if the D-Wave machines do strut their quantum stuff?
Even if they do eventually demonstrate a quantum speed-up, they won’t be as good as fully quantum computers, which are still at least 15 to 25 years away, says Smelyanskiy. He says the comparison is similar to the way Charles Babbage’s 19th century analogue difference engine, the precursor to today’s computers, measures up with your current PC. At the moment, D-Wave’s machine is more like a rusty difference engine that doesn’t seem to work properly, but if D-Wave can clean it up, the results could be impressive. “It will probably have the same impact on mankind as the real difference engine had,” he says.
When can I have my own personal quantum computer?
Whatever the future of quantum computers, don’t expect to own one yourself. “This will be a special-purpose device that can solve a limited set of problems much better than a classical one, but it will never be a general purposes machine like your laptop or your iPhone,” says Troyer. “It’s not what we’ll have at home in the future.”