DWave's quantum processor.

[NanoStuff]

http://dwave.wordpress.com/files/2007/01/sample-holder-with-europa-chip.JPG

Looks sleek, can't wait until motherboards have sockets for this baby

 

[Cheex]

I'm not stupid but I just have to ask....

What kind of CPU is that?

 

[soydios]

The little black square in the middle looks like the die. It seems to need the traces going outward to the pins because it's so small, since it's a quantum processor.

 

[NanoStuff]

Originally posted by: Cheex
I'm not stupid but I just have to ask....

What kind of CPU is that?

Like I said, it's a quantum CPU The only thing I could gather is that it's a 16 qbit adiabatic on a superconducting 'substrate'. They seem to be pretty tight lipped about the details. It's gonna go on show later this year, so more information might pop up.

They claim 1000 qbits by 2008, which would be unreal if true, that's more than enough for mind boggling calculations.

 

[Cheex]

What is a qbit?

 

[NanoStuff]

Originally posted by: Cheex
What is a qbit?

It's like a bit (transistor that is either on or off). It can likewise be on or off, but it can also be on and off at the same time, which leads to various entanglements with parallel universes and some such duality dead cat nonsense. In other words, nobody really knows

 

[pm]

After a bit of googling.
The company: http://www.dwavesys.com/

They claim a public demo in Feb.

A blog about it. http://dwave.wordpress.com/
It has photos of the die (which looks extremely sparse to say the least... and on really old litho it looks like). It is definitely not super small - at least the litho doesn't appear to be - the resolution, the way die photo is imaged... At a guess, this isn't 45nm or even 90nm... it's much, much bigger. I'm going to guess that it's well above 1 micron based on the image.

A paper about the method used - at at least the one that appears to be used:
http://arxiv.org/PS_cache/cond-mat/pdf/0608/0608253.pdf

Based on the paper - which is on a AQC design as well - it's a silicon process with niobium as the superconductor. The paper describes the device being chilled to 0.01 degrees Kelvin. You program the thing by using tunable couplers. It appears that you read the result by examing the coupler flux bias. I'm not sure how you would measure this, but I'm going to guess it's not VGA out. Or even DVI.


Found this: http://www.csr.umd.edu/csrpage/research/quantum/index.htm
It's a somewhat readable description of how these things work.

 

[gerwen]

Originally posted by: Cheex
What is a qbit?

Old length unit. Roughly the length of the average forearm

Wiki qbit

I heard the new processor gets forty rods to the hogshead, and that's they way they likes it.

 

[StopSign]

Originally posted by: pm
The paper describes the device being chilled to 0.01 degrees Kelvin.

Nice! Maybe this cooler can bring NetBurst to 10 Ghz.