How Quantum technology could transform computing - perhaps…

There's still some time - and a number of challenges to overcome - before quantum computing becomes truly useful, reports Nic Fearn

When Google claimed that its latest quantum computer had achieved ‘quantum supremacy', the claims were initially reported somewhat uncritically.

The company had claimed that its device was capable of quickly performing a particular calculation that conventional computers would take years to perform - in this case, Google argued, 10,000 years.

But after bathing in pages of adulatory press coverage for the best part of a month, Google's main rival in quantum computing, IBM, hit back.

They are more or less useless for number crunching and anything demanding 'right-first-time' precision

Google, claimed IBM researchers Edwin Pednault, John Gunnels and Jay Gambetta, hadn't achieved quantum supremacy at all. Rather, their own simulation indicating that a conventional computer would take 10,000 years to perform the particular calculation was, not to put too fine a point on it, wrong.

In order to be able to make that claim, they had made some erroneous assumptions and omitted a number of conventional computing technologies that should reduce the time required to perform the calculation from 10,000 years to just two-and-a-half days.

The quantum supremacy concept was originally coined by John Preskill, the Richard P. Feynman professor of theoretical physics at the California Institute of Technology, as a test that would indicate that quantum computers were finally able to outperform conventional computers. That point, it seems, is still some way off.

But behind the hype, smoke, dry ice and mirrors of quantum computing today, what is the truth? Is quantum computing set to go mainstream any time soon or does the technology present a number of almost insurmountable challenges preventing it from being more than a curiosity?

"Quantum computers are vital for modelling and understanding chemistry/chemical reactions, biology, life, weather systems, climate change, volcanology, tectonic movements and economics," says Peter Cochrane, professor of sentient systems at the University of Suffolk and former chief technology officer at BT.

To do anything really useful, we need around 100 qubits, and if we could assemble 1000 reliable qubits we would become gods

"Quantum computers are essentially analogue. Some claim pure digital operation, but the error rates are so high quantum computers look to be more analogue than digital. They are more or less useless for number crunching and anything demanding 'right-first-time' precision."

Powerful machines

Quantum computers do have the potential to be incredibly powerful and capable, but how do they actually work? Cochrane explains that stable quantum processes are only possible in low-noise environments, insulated from mechanical vibration; low electrical noise; and, low thermal noise. So they must be insulated, screened and decoupled from the world and operated close to, or less than, 0.01 degrees kelvin (0.01k) - that's -273 degrees celsius.

That, in and of itself, presents a number of major challenges related to the production and wider use of quantum computers, not to mention major issues of energy consumption and, hence, sustainability.

Cochrane continues: "From a standing start at room temperature, it takes about 24 hours to get down to 0.01k. They also need a companion digital computer to perform error correction and to select the most likely statistically correct answer, and therefore the amount of computing energy is not generally large."

In other words, qubits as a proxy for how powerful a quantum computer is can be somewhat misleading, because so much of the compute power quantum computers supposedly provide have to be devoted purely to error correction.

Compared to traditional computers, these machines also have specific cooling requirements.

"The cryogenic plant required is extensive to keep a quantum computer stable to within a 0.001k variance. This is also the biggest consumption of energy with a quantum computer typically demanding about 30KW for cooling," says Cochrane.

How Quantum technology could transform computing - perhaps…

There's still some time - and a number of challenges to overcome - before quantum computing becomes truly useful, reports Nic Fearn

When it comes to building a quantum computer, Cochrane says that current models predominantly use silicon. But in the future, materials for simple qubit devices may include aluminium or niobium. He says: "Both are superconducting at cryogenic temperatures and may be integrated with electron-tunnelling devices."

Cochrane describes the manufacturing process as highly specialised, saying that there are many configurations for a qubit, adding that the structures demand great precision and repeatability. He explains: "A complete quantum computer has to have all its qubits within a diameter of about 15cm or quantum coherence is lost.

Much of the compute power quantum computers supposedly provide have to be devoted to error correction

"The maximum number of qubit claimed today is around 50, and most are working with 22 or fewer. To do anything really useful, we need around 100 qubits, and if we could assemble 1,000 reliable qubits we would become gods.

"There are a very limited quantum computer algorithms today, but the reality is that no quantum computer to date has got close to the performance of a top end PC, mediocre mainframe or a supercomputer. As of today, quantum computers are factorising numbers like, 15, 21 and 63."

On top of that, due to the risk of results being corrupted by noise, algorithms need to be as simplified as possible to prevent ‘decoherence' and provide reliably accurate results.

Professor Andrew Lord, head of optics research at BT Labs in Martlesham, Suffolk, notes how quantum computers are already becoming a reality. He says: "Today, the largest known gate-model quantum computer has 72 qubits (that's Google Bristlecone) and the largest quantum known annealer has 5,000 qubits (DWave Pegasus). In both cases, this is the number of physical qubits, but this is equivalent to a smaller number of idealised ‘logical qubits' after error correction."

Looking ahead

Quantum computers, when they become powerful enough, offer a great deal of potential for solving the kinds of problems beyond the capabilities of conventional computing.

A complete quantum computer has to have all its qubits within a diameter of about 15cm or quantum coherence is lost

Tim Callan, senior fellow at certificate authority Sectigo, describes the area as a different computing paradigm. He says that because quantum computing is not the traditional set of discrete 1/0 gates that govern traditional chip technology, it can have a great advantage over traditional computing for specific tasks.

Callan says: "One of those is anything that involves factoring large prime numbers. This fact matters because the cryptographic algorithms used throughout our global digital infrastructure depend on the difficulty of calculating primes."

In other words, quantum computing also has the potential to make cracking the strongest encryption available today a thoroughly trivial task.

Callan believes that quantum computing will also eventually reduce the time required for a brute force attack against deployed cryptographic schemes by orders of magnitude. Callan adds: "As that happens, the basic security that protects our financial systems, commerce, communication, transportation, manufacturing, supply chains, government, and all other aspects of digital life will cease to be effective."

Jason Soroko, chief technology officer at security firm Sectigo, expects these computers to become even more powerful over the coming years. "Quantum computers are typically measured in the number of stable qubits. IonQ claims to have a quantum computer with 79 processing qubits. This is more than Google's 72 qubit Bristlecone processor," he says.

It has been said that a single 100 qubit quantum computer could be more powerful than all of the classical supercomputers combined

"It is estimated that within the next decade, quantum computing will be at a point where it can threaten cryptographic algorithms that are currently commonplace, such as RSA 2048 and ECC 256.

"It has been said that a single 100 qubit quantum computer could be more powerful than all of the classical supercomputers combined. This would [only] be true for specific applications, such as artificial intelligence, pattern recognition and other tasks where classical computers are slow."

Dr Kevin Curran, a senior member at the Institute of Electrical and Electronics Engineers (IEEE), believes that quantum technologies will lead to faster and more accurate computation in many aspects of computing. "For instance, in finance, they could consider millions of individual investment scenarios and calculate which have the best chance of long term success," he suggests.

But Todd Moore, vice president of encryption products for cloud protection and licensing activity at Thales, is more sceptical. He says that although there are some good examples of basic quantum computing today, it still has a long way to go before it becomes a reality. "Quantum computers tend to be slow and limited in what they can do," says Moore.

However, he adds, that as the advancement of quantum computing processing power gathers pace it will also send shockwaves around the world, as they render the public key cryptography that most organisations use today useless.

Moore explains: "In fact, the cryptography that organisations deploy for everything from ensuring that credit cards, personal identifiable information, intellectual property and other sensitive data are secure will be easily exposed by quantum computing - a very lucrative tool for hackers should they manage to leverage it."

And those hackers will almost certainly be working for nation states towards political goals.

In the meantime, though, quantum computing remains some time from everyday commercial use, if ever, due to a combination of the difficulty of building even modestly powerful machines combined with cost.

That will leave quantum computing a curiosity very much in the hands of just a few well-resourced national governments, institutes and major companies like Google and IBM. And it remains to be seen whether some of the impracticalities will ever be overcome.