This is just mind boggling. Quantum related things is one of the reasons I decided to go back to college. I'm just lost in a vortex of unknowing. Props to IBM research team for being able to find the errors within the system.
The D-Wave is not a quantum computer by the usual meaning of that term.
It's only a "quantum computer" by a definition created by their market department. A definition that also classifies as "quantum computer" the one you have on your desk.
That's not entirely fair. D wave is attempting to use quantum effects anneal to solutions faster than a classical computer would. It's not what we would usually think of as a quantum computer, but its in the right ball park. No?
They claim to make use of quantum effects -- in an abstract sense, quantum tunnelling across potential peaks in the solution space to access classically unreachable absolute minima. If that claim is true, then I agree that D-Wave is a type of quantum computer. What is disputed, I believe, is the truth of the claim.
Well my desktop also uses quantum effects. So did valve computers at the time, thus, I don't think anybody ever built a "classical computer".
Yes, that's the definition they use on their site. No, I don't think applying it verbatim to my computer is unfair.
Ok, they created a different kind of classical computer. One that may be able to solve some kinds of problems better, like a specialized computer often does. Yes, it could have been huge (but wasn't - too bad). But it's not a quantum computer, and their marketing is dishonest.
>If a quantum computer could be built with just 50 quantum bits (qubits), no combination of today's TOP500 supercomputers could successfully outperform it.
I would really, really like to know how they came up with this.
2^50 is the upper bound on the size of the solution space that can be searched by a 50-bit quantum computer. If that same solution space were divided up amongst the hundreds of thousands of cores in the world's largest supercomputers, the solution could be found in much less than 1 second, or indeed much less than 1 cycle of any computer, quantum or otherwise.
Anyway, whatever "cycle" means in a quantum computer, it would take much more than 1 cycle to search the whole solution space. Assuming your problem is in BQP you still have to iterate until error is within acceptable limits.
16 comments
[ 2.8 ms ] story [ 52.2 ms ] threadhttp://cdn.phys.org/newman/gfx/news/hires/2015/3-scientistsa...
It's only a "quantum computer" by a definition created by their market department. A definition that also classifies as "quantum computer" the one you have on your desk.
Yes, that's the definition they use on their site. No, I don't think applying it verbatim to my computer is unfair.
Ok, they created a different kind of classical computer. One that may be able to solve some kinds of problems better, like a specialized computer often does. Yes, it could have been huge (but wasn't - too bad). But it's not a quantum computer, and their marketing is dishonest.
I would really, really like to know how they came up with this.
calculating with 2^50 states at once in one cycle certainly outperforms any known super computer, even if a cycle costs 1s. It does not however.
Anyway, whatever "cycle" means in a quantum computer, it would take much more than 1 cycle to search the whole solution space. Assuming your problem is in BQP you still have to iterate until error is within acceptable limits.