Skip to Content
Uncategorized

D-Wave’s Quantum Computer Goes to the Races, Wins

Tests suggest that a CIA-backed quantum computing technology can be very powerful for some kinds of problems.

When I visited D-Wave last year I saw some spectacular hardware and heard of significant backing for the company (see “The CIA and Jeff Bezos Bet On Quantum Computing”). But no one was able to show me results from pitting one of D-Wave’s unusual computers directly against a conventional one to prove how much faster they could be.

The chip at the heart of D-Wave’s computers is cooled close to 0.02 Kelvin, colder than anything in the known universe. Credit: D-Wave

Now the first results of a proper race between D-Wave’s quantum machine and conventional software on a regular computer have come out, and they look good for the Vancouver company (although they don’t address the question of whether quantum or conventional physics are what makes its chips work).

Catherine McGeoch, a computer science professor at Amherst College, carried out the tests and will soon present her results in a peer reviewed paper at the International Conference on Computing Frontiers. Her verdict on D-Wave’s computer? “In some cases, really, really fast.”

McGeoch is an expert in “experimental algorithmics” – algorithm racing, essentially – and conducted her tests using three examples of what are known as “optimization” problems. These are the mathematical core of conundrums such as figuring out the most efficient delivery route around a city, or how the atoms in a protein will move around when it meets a drug compound. In each trial, she pitted a D-Wave computer, a giant black cabinet hiding a chip cooled to 0.02 Kelvin, against software running on a Lenovo workstation with a 2.4GHz quad core Intel processor and 16GB RAM.

Some of the results saw D-Wave’s machine win in spectacular fashion. On one problem well-matched to the hard-wired design of the machine’s super-cooled chip, it found the best result about 3,600 times more quickly than the best conventional software solver. It crossed the finish line in just under half a second, while the second finisher took 30 minutes.

However, another contest saw D-Wave’s machine do less well. It was of a type that required helper software to be used to translate the problem posed into a form suited to D-Wave’s chip, and the quantum system performed much the same as the conventional software and computers.

Each of these cabinets is one D-Wave computer. Credit: D-Wave

In a third trial, which also required use of D-Wave’s helper software, the conventional computers lost again, but not too badly. The D-Wave system found 28 of 33 solutions given 30 minutes to work on the problem, with the closest conventional software finding only 9.

McGeoch says her results are just the beginning of the work needed to figure out how capable D-Wave’s computer are, but sums up her findings like this:

“This type of computer is not intended for surfing the internet, but it does solve this narrow but important type of problem really, really fast…If you want it to solve the exact problem it’s built to solve, at the problem sizes I tested, it’s thousands of times faster than anything I’m aware of. If you want it to solve more general problems of that size, I would say it competes – it does as well as some of the best things I’ve looked at.”

McGeoch says that could prove tempting to large companies such as Google, or government agencies.

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

Google’s Gemini is now in everything. Here’s how you can try it out.

Gmail, Docs, and more will now come with Gemini baked in. But Europeans will have to wait before they can download the app.

This baby with a head camera helped teach an AI how kids learn language

A neural network trained on the experiences of a single young child managed to learn one of the core components of language: how to match words to the objects they represent.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.