The future of computing

Engineers make quantum devices at the National Fabrication Facility at the University of New South Wales

In 1943, IBM’s first president, Thomas J. Watson, allegedly made this spectacularly incorrect prediction: “I think there’s a world market for maybe five computers.” In recent times, the authenticity of this quote has been contested as some believe Watson – if he did say it – was referring to adding machines the size of a house powered by large vacuum tubes.

Three years later, physicist, Sir Charles Darwin, who was director of the British National Physical Laboratory and grandson of the famous 19th century naturalist with the same name, wrote: “It is very possible that ... one machine would suffice to solve all the problems that are demanded of it from the whole country.”

Darwin too was terribly wrong. Regardless, there was simply no way either man could have predicted the tsunami of modern silicon chip-based servers, desktop PCs and mobile devices, such as the modern Apple iPhone, that would flood the world over the next 70 years.

And what a mighty wave it has been. In June 2008, Gartner predicted that there would be two billion PCs in use worldwide this year.

Now the world of computing is once again on the cusp of a revolution. And computers using quantum and nano technologies, in particular, may lead the way when microprocessor manufacturers finally reach their physical limits.

Quantum systems are based on the principles of quantum physics and work by generating complex calculations at the atomic and sub-atomic level. They promise a future where we have access to computational power well beyond the current capabilities of a traditional PC or supercomputer.

These machines have the potential to impact everything from modelling complex drugs to perhaps even one day unravelling the secrets of the universe. But the question is: How far off are we from having quantum systems available for commercial use?

Making it happen

Canadian firm, D-Wave Systems claims it has already commercialised a system using a subset of quantum mechanics called quantum annealing. Yet in February, researchers at the University of California and IBM questioned whether a D-Wave machine used by Google actually relies on quantum mechanics.

However long it takes for mass-produced quantum computers to become a commercial reality, academics at universities in Australia and abroad are leading the charge with their quantum discoveries.

Last November, researchers at the University of Sydney, the University of Tokyo and the Australian National University created the world’s largest quantum circuit board, an essential component in high-powered laser light computers.

During the experiment – which was proposed by Dr Nicolas Menicucci, a theoretical physicist from the University of Sydney’s School of Physics, and conducted at the University of Tokyo – 10,000 quantum systems were brought together in a single component.

Just over a year earlier, researchers at the University of New South Wales created the world’s first working quantum bit (qubit) based on a single atom in silicon. They say this will lead to the development of ultra-powerful computers in the future.

The research team was able to read and write information using the ‘spin’ or magnetic orientation of an electron bound to a single phosphorous atom embedded in a silicon chip. A month earlier, the team created a single-atom transistor, which they believe could be used as a building block for quantum computing.

Another team of researchers at the Australian National University, the National University of Singapore, and the University of Queensland suggested that background interference in quantum-level measurements – known as quantum discord – may be the key to unlocking quantum computing’s potential.

These are just a few examples of the research into future computing taking place worldwide. So why is there such a drive around quantum research when classical computers are still doing the job?

“Quantum computers are not yet at the level where they are doing problems that are too hard for a classical computer,” says the University of Sydney’s Dr Menicucci. “The regular computers we have currently outperform all existing quantum computers on tasks that have been done so far.”

But what quantum computers do have is the potential in the long-term to solve complex problems or algorithms that need to be massively scaled. “There are certain computing tasks that get much harder as you scale them up,” says Menicucci.

“Multiplying numbers is easy when it’s three times two. As the numbers get bigger, you may have to use a calculator, but even huge numbers are no problem; multiplication is relatively easy.

“But factoring whatever number you get out of that multiplication into the original numbers is a problem that our best algorithms are not very good at. As the number to be factored gets bigger, it becomes astronomically difficult to do. You can easily generate a modest-sized number that is so hard to factor, it would take the best supercomputer in the world longer than the age of the universe to factor it.”

But this is not true of a quantum computer, says Dr Menicucci. “One of the things that a quantum computer can do very quickly, if we could build one, is factor these very large numbers. We don’t know of a way a classical computer can do that in an efficient way if the number gets big.”

Everything is quantum

Everything in the world is quantum – all matter, all energy obeys quantum mechanics, continues Dr Menicucci. “The question is: ‘Why don’t we see these quantum effects in our daily life? Why do we have to work so hard to engineer them?’”

The answer, as Dr Menicucci explains, is that you need “very fragile states” to see quantum effects. “Think about it in terms of a deck of playing cards – if you take out the cards, throw them on a table randomly and someone walks by and creates a small breeze, nothing happens to the cards.

“If someone sneezes, maybe they move a little bit but not much really changes. But if you take those cards and carefully arrange them into a pyramid and someone walks by, bumps the table or sneezes, it’s all gone.”

One configuration of the cards is robust and the other is fragile, Dr Menicucci says. “Quantum physics is like that in the sense that quantum computers need very fragile quantum states – known as entangled states – to do their calculations.”

Despite the obvious difficulties of keeping fragile quantum states together, research around information processing is showing real promise. But researchers need new theoretical developments and experimental breakthroughs to make it work.

“The first ordinary computers were room size, inefficient and broke down,” Dr Menicucci says. “We’re not even at that level yet, we’re at the level of just designing and making robust the individual ‘vacuum tubes’ of a quantum computer, if you will – and testing them in small quantum devices.”

According to Dr Menicucci, a lot of work remains to be done to scale these quantum machines to be capable of tackling real-world problems. “The rest of the world is walking by our house of cards, sneezing on it and knocking it over. That’s a phenomenon called ‘decoherence’… and it shows up as noise that ruins your computations,” he says.

“Figuring out clever ways to fight decoherence is one of the main things that both theorists and experimentalists are working on right now.”

Next up: Leading the way

Page Break

Leading the way

For conventional computing, silicon is still hard to beat as a billion transistors can be placed on a single chip, says Dr Andrew Dzurak, a Scientia professor in nanoelectronics at the University of New South Wales and director of the Australian National Fabrication Facility. Dr Dzurak was part of the team that created the qubit based on the nucleus of a single atom in silicon.

“At the moment in the world of making quantum transistors, or quantum bits, it’s not so cut and dried,” he says. “When you are working with quantum bits that can be single atoms or single molecules or small physical systems, there are a lot more options at play.

“My own research is in silicon and I can argue good reasons why silicon is a strong choice for quantum computing as well. But there are other candidates; promising ones that need to be considered in the quantum mix.”

One of these is ‘ion traps’, a technology used in atomic clocks. An ion trap employs magnetic and electric fields to capture ions in a vacuum chamber. Quantum information can be stored on a single ion and then accessed using laser beams.

“Researchers have put together a number of these ions and performed rudiamentary calculations. Each one of these ions is a charged atom with an electron stripped off, and they can act as a quantum bit,” explains Dr Dzurak.

“One of the aims of all of the groups working on quantum computing is to scale up from one quantum bit to many. Ultimately, we’d like many thousands or even millions [of quantum bits], and in terms of the most advanced quantum technologies at the moment, ion traps currently lead the way.”

Another strong candidate is superconductor materials such as Niobium, says Dr Dzurak. Several groups worldwide are also working on creating superconducting circuits, where each small circuit or device can act as a quantum bit.

UNSW’s discovery of a qubit on a single atom in silicon – where information can be stored and read using the spin of an electron bound to a phosphorous atom in a silicon chip – is also a promising breakthrough.

“There are other materials you can also use to store information on the spin and one of those is diamond,” says Dr Dzurak. “There’s a naturally occurring defect in the crystal structure of diamond with associated electrons that have a spin. That’s another interesting prospect for quantum computing, or perhaps in that case, for quantum communications.

“Groups around the world are exploring different possible realisations for quantum bits and it’s a race.”

Industries to benefit

The semiconductor was invented at Bell Laboratories in the US in 1947 and it was many years before silicon chip-based computers started to outperform other computing systems.

As Dr Dzurak points out, it took our current computing industry decades to go from inventing the first transistor, to providing commercial systems.

Transistors were in fact first used in hearing aids, which provided a commercial driver to move the technology forward. The quantum computing world is now looking for an equivalent small-scale commercial application that can provide momentum while it strives towards the long-term goal of combining thousands of qubits together and building a system that can outperform existing computers.

“Having a near-term commercial problem to solve will help a great deal,” says Dr Dzurak. “This doesn’t mean things won’t get done though. There are certainly investments from around the world by national governments and research agencies that realise quantum computing is something to work towards.”

Regardless of which type of technology wins in the end, future quantum computers may solve a broad range of problems across several industries. Pharmaceutical firms, for example, could use these systems to model or simulate molecules or structures for new medicines.

“Existing computers are not able to solve those problems; once a molecule gets larger than 20 or so atoms that make it up, it becomes computationally too difficult for even our largest supercomputers to solve,” says Dr Dzurak. “That’s where quantum computers will be ideally suited to simulate these types of natural physical systems.

“A pharmaceutical designer could sit in front of their computer, move atoms around and try out ideas – for example how molecules interact with each other.

"That’s a wonderful dream to have, to be able to design medicines on a computer rather than to run countless medical trials for years and years. Of course the final drug would still need a proper trial, but many years could be saved in the process.”

The manufacturing industry may also use quantum computers for complex calculations that could help engineers design new lightweight materials that may, for example, enable aircraft to fly longer distances using less energy, says Dr Dzurak.

There may also be applications in the field of artificial intelligence. It’s possible quantum computers could simulate the complex processes that occur inside the human brain, he says, although he emphasises that this is just speculating.

“There’s a very good likelihood that within five years a small-scale quantum computing system would be able to solve a particular problem that currently can’t be solved with the biggest super computer on the planet,” Dr Dzurak concludes.

The first quantum machines

D-Wave has built what it claims to be the first and only commercially available quantum computers already being used by a range of organisations including Lockheed Martin.

The American aerospace and defence firm uses a D-Wave quantum machine to speed up software verification and validation of its flight control systems. “These are incredibly complex systems and testing the software is often the largest optimisation problem in a new aircraft project,” says D-Wave’s CEO Vern Brownell. “Lockheed has been using a D-Wave system to find faster, better ways to do this.”

NASA is also using the manufacturer’s quantum technology to search for new planets outside our solar system that might harbour life. Google, meanwhile, is exploring machine learning to build more accurate models for everything from speech recognition to bioinformatics, Brownell says.

“The D-Wave is best suited to tackle challenges in big data analytics, machine learning and complex optimisation problems – all of which exist in many different domains such as mission planning, pattern recognition and anomaly detection, cancer research and finance,” he claims.

One of the company’s goals is to make quantum computing accessible more broadly through the cloud. “Quantum computers can complement traditional computers and we envision a model where problems best solved by a quantum computer will be sent through the cloud to a system a returned to the user,” Brownell says.

An important part of D-Wave’s development is its roadmap, which has been created over the past decade. According to Brownell, the company’s systems are matching and in some cases surpassing the performance of state-of-the-art classical computers that have been painstakingly developed for over 60 years and reaped trillions of dollars in investment.

“We continue to refine and innovate our processor and are planning to release the next 1000 qubit processor later this year,” he adds.