Tech frontiers
IT PROMISES TO TRANSFORM THE WAY WE LIVE, BUT WHEN WILL WE SEE IT? WE LOOK AT THE HEAD-SPINNING FUTURE TECH THAT IS QUANTUM COMPUTING.
IT’S ARGUABLY THE most complex theoretical computing concept to try to wrap your head around, but governments, universities and big business are throwing serious dollars at trying to make it a reality. It promises untold potential and the ability to quickly solve mathematical problems that would take current computers years to figure out. But to understand the brain-spinning concept of quantum computing, it’s important to remember just where we’ve come from.
If you were lucky this Christmas, you woke up to find a NES Classic Edition games console under the tree. While it looks to be a brilliant way to scratch your classic-gaming itch, it’s also the latest example of what’s known as ‘classical computing’. Your PC, games console, your phone or tablet — they’re all examples of classical computing, relying on the basic tenant that a switch has two positions or ‘steady states’ — on and off. In digital parlance, those states are given the digits ‘1’ and ‘0’ to form one ‘binary digit’ or ‘bit’ for short.
As humans, we’re used to thinking in tens, using digits 0 to 9, but computers count in twos or ‘binary’, using just ones and zeros. Combine two switches together and you have two bits that can cover any of four combinations — we can write it as 00, 01, 10 and 11. Make it three switches and you now have eight possible combinations — 000, 001, 010, 011, 100, 101, 110 and 111. From that, we can use a more mathematical way of describing it as ‘n’ bits can represent ‘2 to the power of n’ states, where ‘n’ is any positive number you like. Our digital devices use electronic switches called ‘transistors’ — and not just two or three of them. The Apple A10 chip inside your iPhone 7 reportedly has 3.3 billion of them, all grouped in various combinations for different functions.
WHY QUANTUM COMPUTING IS DIFFERENT
Quantum computing also uses the concept of bits, but here, they’re called ‘quantum bits’ or ‘qubits’ for short and they turn the classical bit on its head. If you think of a classical bit as a straight line on a piece of paper, with ‘0’ on one end and ‘1’ on the other, it can only be in one of those two states — there’s no middle ground.
However, a qubit can be both zero and one at the same time. That’s probably enough to do your head in right there, but a helpful way of visualising a qubit is to make that straight line a centre vertical or ‘Z’ axis and spin a sphere around it, with the ‘North Pole’ as zero and ‘South Pole’ as one. Often called a ‘Bloch Sphere’ and named after Swiss physicist Felix Bloch, any point on the surface of that sphere is a possible location or ‘superposition’ that can be represented by a qubit and expressed as a mathematical function of the two poles. But if a qubit is capable of being both zero and one simultaneously, it means you can now have ‘n’ qubits representing ‘2 to the power of n’ superpositions at the same time. Meanwhile, the poor old classical bit can only be one of its ‘2 to the power of n’ values at once. This gives qubits excellent ‘parallelism’, meaning the possibility of performing many calculations at the same time.
HOW FAST IS A QC?
With parallelism comes the potential for incredible processing speed, one of the factors driving the dream of quantum computing over much of the last 40 years. Performance numbers at the moment are mostly theoretical, but the common example floating around academia for the last ten years or so is that a 30-qubit quantum computer could theoretically deliver the same processing power as a modern computer running at 10 teraflops or 10 trillion floating-point operations per second. Finding a comparable modern-day equivalent in terms of ‘flops’ is fraught, but if you consider a Sony PlayStation 4 games console has a graphics chip with a claimed peak at 1.84 teraflops, you get a very rough idea of where things sit. And given some of the numbers being thrown around, 30 qubits isn’t that many.
CONTROLLING ATOMS
But making quantum computing a reality has proven tricky for a number of reasons, not least of which is that the technology is based on the concept of controlling the orientation of angular momentum or ‘spin’ of an individual electron bound to a single atom in a magnetic field. By altering the orientation of an electron’s spin, whether ‘up’ or ‘down’, the electron itself can be made to store data and function as a qubit. The problem is, keeping a single electron in its ‘quantum state’ long enough to perform complex calculations has proven exceedingly difficult. This phenomenon is called ‘coherence time’, and up until recently, the amount of time an electron could be made to continue its controlled spin was measured in microseconds.
THE RESEARCH
The University of New South Wales (UNSW)
achieved a major breakthrough in 2013, with researchers able to control the spin of a single phosphorus atom’s core or ‘nucleus’ for an impressive 60 milliseconds. It doesn’t sound very long, but imagine a smartphone battery that lasts you a day now lasting you 600 days and you get the idea. By 2014, coherence times of beyond 30 seconds were being discussed.
In 2015, the university made another breakthrough when researchers succeeded in building a two-qubit logic gate in silicon, the same stuff current computer chips are made from. It was a huge development, not just for being the first time calculations between two qubits had been performed in silicon, but as silicon is the ‘bread and butter’ material of the electronics industry, any future production should greatly reduce manufacturing costs.
More recently, UNSW researchers announced in October 2016 they had developed a more robust way to encode the spin of that phosphorous atom and achieve a ten-times improvement in coherence. But it’s not just universities who are neck-deep in research. It shouldn’t surprise (but somehow still does) when looking into new technologies of the future just how often many of the same familiar names pop up — names like Microsoft, IBM and Google.
Microsoft has had a base-camp set up at the University of California since 2005 called ‘Station Q’ to research the potential of ‘topological quantum computing’, which, very simply, is encoding qubits using pseudoparticles called ‘anyons’ made by splitting electrons. The company renewed its pledge in November 2016 to develop practical, scalable quantum computing hardware and software, with news that long-time Microsoft engineering manager Todd Holmdahl, who previously drove Xbox and HoloLens development, has taken the reins of its quantum efforts. The software giant is also joining with select universities from around the world to boost research. One of those universities is the University of Sydney, which, in 2016, launched the Australian Institute of Nanoscale Science and Technology to drive research into nanoscience as a whole.
THE CODE
However, hardware is only half the problem — the other half is software and, right now, there’s no definitive answer on how to best program a quantum computer. As development continues, there’s a proliferation of quantum computing simulations hitting the web in different forms you can have a play with that make quantum computing more practical. One example made by a few Google engineers is the Quantum Computing Playground ( quantumplayground.net). It’s a quantum computer simulation, featuring what engineers call a ‘single quantum register’ that you can program via a scripting language called ‘QScript’. It seems reasonably easy to learn (its a little bit like Python) and the online editor allows you to write, save and run your own quantum computing code. One of its highlighted examples is simple number factoring, using Shor’s algorithm.
The UK’s University of Bristol has come up with a more graphical ‘quantum in the cloud’ quantum processor simulation using single photons ( cnotmz.appspot.com). The idea is that you practice with this as a forerunner to seeking permission to use the university’s actual quantum computing chip.
In May 2016, computing giant IBM launched the Quantum Experience ( tinyurl.
com/ztsrt5v, sign-up required), giving you the coding keys to one of its older five-qubit quantum computing engines, again starting on a graphic-based simulator before moving up to the real thing.
And not to be outdone, Microsoft has also released a quantum simulator called ‘LIQUi|>’ (Liquid), which you can download from its Github site at stationq.github.io/Liquid.
It combines a programming language with a quantum device compiler to convert high-level code into quantum machine instructions and is built on top of Microsoft’s Visual Studio development suite.
WHY THE HURRY?
But what is it about quantum computing that has everyone running around in a frenzy? One possible reason is ‘cryptanalysis’ or code-breaking. Remember the 1992 hit movie ‘Sneakers’ with Robert Redford and Ben Kingsley chasing after a black box codenamed ‘Setec Astronomy’ that could break every code? Quantum computing could very well end up being the ultimate code-breaker, so it’s not surprising everyone wants to get their hands on one. “Too many secrets” indeed!
Here’s a brief look at why. In our recent feature, ‘Ten scientists who changed the tech world’ (December 2016, page 74), we included British security pioneer and mathematician Clifford Cocks. He independently came up with a cryptosystem in 1973 known today as ‘RSA’, one of the key algorithm sets protecting the internet (the British Government kept Cocks’ discovery classified for more than 20 years — it was later patented by Rivest, Shamir and Adleman at the Massachusetts Institute of Technology, who independently developed it in 1977. It was only released to public domain in 2000). It works on the concept of two keys — a public key and a private key. Part of both keys is the product (multiplication) of two very large integer prime numbers (numbers that are divisible by only themselves and one). Both keys also contain separate coding factors — the encoding factor in the public key, the decoding factor in the private key. The public key can be divulged along with the coded information because, without the private key, the coded info can’t be deciphered — well, at least, not easily.
To crack the code without the private key, you’d have to try to work out the ‘prime factors’ of the public-key number. Factoring is what you do when you take a number and try to find the prime numbers or ‘factors’ that, when multiplied together, give you that original number. For example, if looking for the factors of 35, you’d want 5 and 7.
However, with so much computing horsepower available in the cloud, security experts are concerned about security of keys that are 512-bits long (think 2 to the power of 512). Cracking keys of this magnitude still takes massive amounts of computer power to find the two prime numbers that made up the key, but in October 2015, researchers developed a method that potentially allows anyone to crack a 512-bit RSA key using Amazon cloud computing at a cost of $75 and four hours. In a nod to the current craze in cloud services, the researchers dubbed it ‘factoring as a service’ (
[ WWW.TECHLIFE.NET ]
But quantum computing, on the other hand, will potentially make very light work of factorisation, which is one reason why countries are pouring significant dollars into making quantum computing a reality.
Of course, it’ll also have more altruistic uses, particularly in medicine and engineering, from development of new drugs to creation of more-accurate weather forecasts. It’ll also likely be deployed against huge databases and even further aid improvements in artificial intelligence. But it’s not hard to see digital security driving the technology forward — for example, the Australian and NSW governments, the Commonwealth Bank and Telstra have all reportedly stumped up a combined $70 million to create a consortium to develop and commercialise UNSW’s quantum computing research.
THE FUTURE
Bottom-line, quantum computing is so fundamentally different to our current computational methods that it’s impossible to predict just how it will eventually affect us. It’ll require new methods for manufacture and coding, but combined with its high-speed complex processing capabilities, it could help solve the riddle of many of our most pressing problems, from cancer to climate change. It could also find mathematical solutions in a heartbeat that would take classical computers years to solve. If history is anything to go by, it’ll likely have plenty of nefarious uses as well. After all, as intelligent as computers are becoming, they are still a reflection of ourselves.
Right now, quantum computing requires harnessing the power of atoms and building it into a computer chip. But despite some recent impressive advances, it could well be 2025 before we start seeing large-scale commercial quantum computers. Looks like those classical bits will have to do us for a while yet.