D-Wave

[Brett Hall]

There's has been some media chatter over the last week about Google's purchase of a "quantum computer".

"Google buys...quantum computer" http://www.cbc.ca/news/technology/story/2013/05/16/technology-d-wave-quantum-computer.html

"Quantum computers are here!" http://www.thepriceofrice.com/2013/05/quantum-computers-are-here.html

"NASA buys into 'quantum' computer" http://www.bbc.co.uk/news/science-environment-22554494

(Arhhhhh....notice the scare quotes in that previous one...)

That BBC article says "The machine does not fit the conventional concept of a quantum computer, but makes use of 

quantum effects".

So....not a quantum computer then? What do they mean by "quantum effects"?

(From http://www.bbc.co.uk/news/science-environment-22554494)

That article, like almost all the media surrounding this story, seems to regard the D-wave as a black box. The article says:

"Unlike standard machines, the D-Wave Two processor appears to make use of an effect called quantum tunnelling." 

Appears? How can something "appear" (from the outside) to use quantum tunnelling?

The end of that article contains this:

"US giant Lockheed Martin earlier this year upgraded its own D-Wave machine to the 512 qubit D-Wave Two."

Now that *can't* be true, can it? 512 qubits?! Without decoherence?

My guess is that this dwave company are great at marketing - their slick website is scant on details (thought it says all 

the right things about it being super cooled and super conducting (presumably to help prevent decoherence):

http://www.dwavesys.com/en/dw_homepage.html

No doubt they have discovered *something* - but my guess is that it's just some sort of better (classical) architecture. Is there

any reason to think they have really constructed 512 qubit quantum computers? And if you are a company who buys one, and it's

called a quantum computer, and it's said to have 512 qubits...could you expect to factor a 100 digit number in milliseconds?

Another article (http://spectrum.ieee.org/tech-talk/computing/hardware/google-and-nasa-buy-a-dwave-computer) 

refers to the machine as a "supposed quantum computer" but then again states that it is a 512 qubit device. Qubit, again.

So is it a genuine quantum computer with 512 qubits, or not?

Here's the kind of information that leads to my skepticism:

Physical Review Letters publishes peer reviewed articles like http://prl.aps.org/abstract/PRL/v108/i13/e130501 by 

Xu, et al (2012) reporting the factoring the number 143 using a 4 qubit chip. Preprint here: http://arxiv.org/abs/1111.3726

I *thought* that was about where we were at. Okay, so it's 2 years later. Maybe they're up to 6. But 512? Woah. With no decoherence?

What does the 512 really mean? 

Should we invest yet?

Brett.

[Elliot Temple]

On May 19, 2013, at 5:04 PM, Brett Hall <brha...@hotmail.com> wrote:

> Should we invest [in D-Wave] yet?

No.

For D-Wave stuff you want to read Scott Aaronson:

http://www.scottaaronson.com/blog/?p=1400

and there are various earlier posts.


-- Elliot Temple
http://elliottemple.com/