Most of us would be happy with a 2x speed-up when buying a new computer, let alone a machine whose performance was trillions of times faster?

Yet that’s the sort of leap forward Microsoft is hoping to achieve with quantum computing: to create a machine able to handle tasks it would take today’s computers eons to complete.

Quantum computers’ potential for incredible performance stems from their ability to exploit quantum phenomena, the counter-intuitive way that matter behaves at an atomic level. While traditional computers represent data as bits, either as zeroes or a ones, quantum computers use qubits, each of which can represent a superposition of both zero and one at the same time. That ability to represent multiple values at once is what lies behind quantum computers’ exponential jump in performance over classical computers.

Microsoft is working on developing these general-purpose quantum computers, which it envisions will one day help the world solve intractable issues.

Describing what might eventually be possible, Krysta Svore, principal research manager in Microsoft Research, said: “Problems that take billions of years classically, we can solve in a matter of days and weeks on our quantum computer.”

“In carbon capture, we can look to find a catalyst to help extract carbon from the atmosphere, to improve and take steps towards combating global warming,” she says.

“In material science we can look to improve clean-energy solutions. We can study materials on the quantum computer. Find a material that superconducts at higher temperatures enabling, say, lossless power transmission.”

The challenges ahead

Unlike quantum systems made by D-Wave systems, which are suited to tackling a limited number of tasks known as optimization problems, Microsoft is aiming to make a general-purpose quantum computer.

However, one of the biggest barriers to making the dream of a quantum computer a reality is the ability to scale up the number of qubits into a workable computer.

Today experimental quantum chips are only available with a limited number of qubits, far below the thousands of qubits that Microsoft wants to scale a future quantum computer up to.

SEE: IT leader’s guide to the future of quantum computing (Tech Pro Research)

To speed up the rate at which a usable quantum computer can be developed, Microsoft is researching topological qubits. The advantage of these type of qubits is they help address a fundamental limitation when trying to scale up to a full-blown quantum computer.

Qubits are incredibly sensitive to interference from energy in the surrounding environment, and this interference can in turn lead to errors in calculations. This need to avoid thermal interference is why Microsoft has to cool its quantum computer down to just 4 Kelvins, around -450F.

Topological qubits are less susceptible to this interference. This robustness reduces the need for dedicated error-correcting qubits, increasing the number of qubits that can be employed to do useful work.

“At Microsoft we think about moving to lower and lower error rates, that is better and better qubits that are robust against noise and error, that have high fidelity, that are going to allow us to scale to large computations to solve the world’s most challenging problems,” said Svore.

“The topological qubit will get us there, and provide a better foundation to allow us to rapidly scale large numbers of really good qubits.”

For those who want a taste of quantum computing today, Microsoft will release a variety of a quantum computing development tools by the end of this year. These include a programming language for writing software to run on quantum computers, which will be integrated into Microsoft’s Visual Studio IDE, and a quantum computer simulator for running quantum computer programs, which can be run locally or on Microsoft’s Azure cloud platform. The requirements will be pretty hefty, with the local version needing 32GB of RAM to simulate 32 qubits.

Further into the future, Svore says to expect that quantum computers will become part and parcel of the modern computing landscape.

“I think we’ll see this amazing range of solutions coming in the next few years, where you’re doing some steps of the computation quantumly and then some classically,” she said.

“These systems will sit in the cloud. Just as today you have options to send your algorithms and applications to a mix of GPUs, CPUs, FPGAs–quantum will be one of those options for you, as you think about accelerating applications.”

Also see