A picture of the demo chip November 9, 2007
Posted by Geordie in QC-Related Posts, Superconducting Electronics, Superconducting Processors, World Domination.trackback
This is what a 28-qubit superconducting AQC looks like. In case you’re wondering, “unholey” refers to the types of transformers between qubits and couplers (no holes in groundplane), “CJC” means this chip used D-Wave compound Josephson junction couplers, and “Leda” is the mask name for this generation. We name our masks alphabetically using moons, so Leda is the 12th generation since we starting design/fab/test of superconducting AQCs. The demo in February was from the Europa mask.
[...] Leda has landed! Dr. Geordie Rose published the first picture of the 12th generation adiabatic superconducting chip by D-Wave [...]
Congrats. I am glad that thing seem to be on track to the large 512 and 1024 qubit systems for 2008. Hopefully you can get the performance that clearly shows the benefits (and the quantum computerness) soon to settle the controversy and enable Dwave to make a lot of dough.
Brian: Thanks! It does look like we’re going to be able to push to 500+ qubits well within 2008, which is extremely cool.
As far as “controversies” go, don’t believe everything you read in the press. The approach to QC we are taking (AQC with superconducting flux qubits) is viewed as being reasonable by the academic community, as is the basic category of quantum algorithms we run on our systems (quantum annealing for solving binary quadratic programs).
Awesome stuff. Keep up the great work, we’re rooting for you
Hi Geordie, things seem to be pretty exciting.
How many qubits do you expect to need before you have a machine that can really blow out of the water any supercomputer (on some specific problem maybe) hands down?
Can you use your quantum computer to make that idiot scott aaronson shut up? then ill believe its a quantum computer
@AIguy: thanks, I’ll see you guys in a few days!
@Jordan: The next generation after this one will have about 500 qubits. I am pretty confident that we will be blowing the competition out of the water for some instances of binary quadratic programs (our target problem) by this generation. So the answer is 500.
@CAM-10: I’m not sure what complexity class that problem is in. I would guess that it’s probably not efficiently solvable by any computer consistent with the laws of physics, but I don’t know of any published bounds.
I really liked your talk at the superconducting workshop a few weeks ago and am glad to see that you are working with people in industry on defining commercial problems for your systems.
I think I finally understand what you are doing, is this correct: (a) take a quadratic unconstrained binary optimization problem (b) embed it into a physical structure where the linear terms are local to the variables and the quadratic terms are local to the structures that couple variables (c) use free energy minimization to drive computation?
Geordie, say hi to the team for me, what you guys have done and are doing is truly amazing… of course you know you owe it all to NG canning TRW SC so you can thank NG’s senior management
This Aaronson guy is a total jackass. On top of which he doesn’t seem to be a real scientist either, take a look at his CV he has maybe 5 total publications in real journals and ZERO physics papers… my theory is that he is being paid by the nsa to discredit foreign qc efforts… and that’s how he got in at mit…sure as hell not based on competence or ethics
@Fred: Thanks, I had fun meeting with everyone in the SCE community (many for the first time). Yes you have it exactly right with the computational procedure.
@Eddy: Thanks!
@Borgo: I think that’s a little harsh… kid’s only 26… were you an MIT prof at 26? Look it’s pretty obvious, because he’s so quotable he’s being manipulated by the press to attract attention to their articles. The appearance of controversy sells.
Do you have plans (in future, when you will have universal AQC) to win $30000 (or maybe more) by factoring RSA-704 (704 bits, 212 decimals digits)?
http://www.rsa.com/rsalabs/node.asp?id=2093
skybio: We do have plans to design processors capable of doing phase estimation for quantum simulation. This type of design should also be able to run Shor’s algorithm. Note that if we succeed in doing this we will be busy fundamentally changing the chemical and pharma industries so we probably won’t have the bandwidth to bust rsa.
Scott Aaronson is indeed a very young individual and completely steeped in academia; two tough hurdles that also create tunnel vision. But he his incredibly bright and certainly has a lot to offer as we step into this strange hyper-cube of science that is bearing down on society. People of his caliber have to be engaged, they must be a part of the conversation.
And Kudos to Geordie and the D-Wave team for this AQC adventure
Could you expand a bit where the chip was fabbed? Was it done in a normal FAB or done by Dwave in-house?
skybio: The factoring contrest has been discontinued, that’s why it’s filed under Historical.
Borgo: I’m not sure where you have been looking, but Scott has a lot more than 5 papers. He’s also extremely competent. I doubt Geordie would argue with that.
I hope that you will be posting the slides and information from your talk(s) at SC07 soon. I am very interested in your views of the theoretical limits and possibities of QC. Unfortunately unable to take the time to be in Reno on thursday.
nork: The chip we’re using for the demo was fabbed at the Jet Propulsion Lab’s microdevices lab in Pasadena.
Brian: Yes I will do this!
[...] Geordie Rose: A picture of the 28-qubit superconducting demo chip [...]
Geordie,
I see the ‘Leda’ but where is the ‘Swan’? By this I mean what problems will this 28-qubit chip be programmed to solve? Do you have interesting problem instances from partners that fit in a 28 qubit encoding?
Other than that I am thrilled that D-Wave has come so far. Well done.
I am curious to know how the demo went last night at SC07. I have been scanning the news and blogosphere, but so far nothing. I hope it went well.
I am wondering what sort of test was conducted? It sounds like you were demonstrating some sort of image recognition algorithm. Are you posting any details about the particular algorithm that was being run?
The main question I have is - how does Orion compare to a classical binary computer in this particular test?
Ooops I see the swan. I missed your previous post with the “application we’re demo-ing is an image matching application co-developed with Dr. Hartmut Neven of Google”.
D-Wave have some project with google?
Geordie:
Congrats on the success of the new chip. Got to tour the lab on Friday, and have to say how cool it was hearing about the success (and the insane amount of work leading up to it!).
D-Wave: Advancing the scientific field of science!
Matthew: Thanks! Yes it was quite the push to get everything working the last few days… the prize is worth the effort!
keep it up, hope it all goes well.
same as 27
First of all, Congratulations!
Questions: What is the time frame (your best estimate will do) for this quantum computer to become commercially available? In layman’s terms, what would be the advantage of owning a such a computer as a desktop (say as compared to the best machines currently available on the market) ? At what number of cubits (provided there would be appropriate A.I. software available) would the machine exceed computational ability of the human brain?
Hi, Geordie.
Our magazine (”Russkiy reporter”, in Russian) is going to print a feature story on quantum computers, and so wee need some high-res pictures of Orion (those already published on the Web but of greater quality). Any official attempts to get them ended by nothing. So if there’s any chance for our photoeditors to contact D-Wave representatives before Monday morning, it would be great.
Thanks in advance.
With best regards,
Borislav Kozlovskii.
When will Linux be ported to this? What about Windows Quantum Edition? Mac OS Q?