The Ongoing Quest for Quantum Supremacy

You might not realise it, but the old era of computing ended about a decade ago. Since the 1960s, we have been able to double the number of transistors that can fit onto an integrated circuit roughly every two years, a phenomenon first noticed by Intel co-founder Gordon Moore. This exponential phase ushered in the information age and transformed communication technology, but all good things must come to an end.

These days, cramming more transistors onto a chip is much more difficult than before. “We are hitting the limits of optical lithography; we can’t make microchips any smaller,” said Associate Professor Dimitris Angelakis, a principal investigator at the Centre for Quantum Technologies (CQT) at the National University of Singapore. “This is because at smaller scales the quantum noise becomes bigger than the signal.”

“On the one hand, quantum effects are part of the problem, but quantum computing is a solution at the same time,” Professor Angelakis told a full-house crowd gathered at 32 Carpenter Street for a dialogue titled ‘Quantum Computers: How They Work and What They’ll Mean for Big Data and Business’. The event was presented by SGInnovate in partnership with CQT and the National Supercomputing Centre Singapore.

The promise of quantum computing lies in the ability to theoretically double the performance of the system with every quantum bit (qubit) added, Professor Angelakis explained. “[A quantum computer] with just 50 qubits would be more powerful than the fastest supercomputer in the world. With 300 qubits, you would have more possible states than the [total number of atoms in the] universe,” he said.

But what would we do with all that power? One application that has captured the public’s imagination is breaking current encryption standards such as the Rivest, Shamir and Adelman (RSA) protocol, which is based on factoring large numbers. Because the RSA protocol is used for everything from bank transactions to virtual private networks (VPNs), it is feared that quantum computing would undermine cybersecurity.

“Quantum algorithms such as Shor’s algorithm would enable RSA encryption to be cracked in just ten hours compared to 10,000,000 CPU years on a classical computer,” Professor Angelakis said. “However, to implement Shor’s algorithm in hardware we would need many qubits, around 4,000. This is not something we can honestly say can be done now or in the near future, so it’s unlikely to be quantum computing’s ‘killer app’.”

As it turns out, the most immediate impact of quantum computing might be felt on something more prosaic: fertiliser. “It sounds a little lame, but five percent of the world’s natural gas goes to fertiliser production."

With a system of a few hundred qubits, Professor Angelakis predicts that quantum computers would be able to address optimisation problems in logistics, finance and engineering, as well as reduce the number of layers in neural networks, giving better results in less time.

At the moment, however, even the most state-of-the-art systems have less than a hundred qubits. In fact, we are still at the first stage where we have to prove that quantum computers are worth all the investment, said Dr Julian Kelly, a research scientist at Google working on the 72-qubit Bristlecone processor.

“If you make a device that can’t compete with a classical computer, why should we spend all this time on it? Proving that there is a real advantage to quantum computers is a milestone we call achieving quantum supremacy,” Dr Kelly said. “The next step would be to demonstrate some near term applications, then of course, in the long term, what we want to do is build a universal fault-tolerant quantum computer.”

Beginning with a nine-qubit prototype developed during his PhD studies, Dr Kelly and his team went on to develop a 22-bit device called ‘Foxtail’ by moving the control and measurement circuitry to the bottom of the chip and stacking two chips together. For Bristlecone, the Google team created a two-dimensional array of six qubits using the same strategy, tiling 12 of the six-qubit units to get a total of 72 qubits.

While 72 qubits are enough to demonstrate quantum supremacy, Dr Kelly was quick to point out that measuring the performance of a quantum computer is not all about the number of qubits. “The important thing to understand when you’re designing quantum systems is that it is all about the errors,” he said. “If the error rate of your qubits is high, your device is going to be useless no matter how many qubits it has.”

“As a result, qubit error mechanisms inform nearly all the design decisions that we make at Google as we build the hardware. They also inform a lot of the limitations that our devices have.”

RELATED VIDEO: Quantum Computer: Hardware and Software Implementations, Challenges and Opportunities

While potential users need to understand these nuances in the capabilities of quantum computers, the researchers working at the leading edge also need to understand the real-world problems that could potentially be solved by quantum computing, said Professor Angelakis. 

“To achieve that, there has to be interaction between quantum researchers and people in industry. We have to educate each other and understand each other’s language,” he said. “That’s why I’m very excited by events like this where SGInnovate acts as the mediator and helps us get together.”

“At CQT, we provide workshops and organise hackathons to allow people to get hands-on experience with our quantum simulators. I’m looking forward to more interactions with people in the various industries.”

We believe that Quantum Computing is the next generation of supercomputers and have recently invested in Archanan and Horizon Quantum Computing. Learn more about our investment philosophy here.