D-Wave Sells First 128 Qubit Computer

[wuliheron]

http://physicsworld.com/blog/2011/05/d-wave_sells_its_first_quantum.html

While laboratories around the world struggle to entangle a few quanta and use them for calculations these people have already managed a working 128 qubit system. To put this in context, a 30 qubit computer would equal the power of a 10 teraflop conventional computer and the power of quantum computing systems goes up factorially with each new qubit added. That's enough power to likely break any conventional encryption systems used worldwide including I suspect that of Wikileaks and most governments.

Europe has been investing in quantum cryptography for quite some time now, especially in the banking sector, which is the only type of cryptographic system theoretically immune to a quantum computer. For its part the US has discouraged even the use of some conventional cryptographic systems in the private sector. The obvious implication is that D-Wave's new computer is the first public shot in a new arms race which will likely last for many decades to come.

 

[bigi]

This is just amazing. Remember reading about quantum computers about 12 years ago, and this was just pure science fiction.

You may get more interest in this finding in Highly Technical forum.

 

[ViRGE]

128 qbits? My knowledge of quantum mechanics is lacking, but shouldn't be possible to instantly break an equally long symmetrical key with a computer like that? E.G. AES-128 just became useless.

 

[dullard]

$10,000,000 for a computer that is outperformed by todays PCs on the one and only one task that it can do? I'll pass for now. Maybe give them 10 more years though and I may change my tune.

 

[wuliheron]

These are qubits, not bits. Like I said, their processing power goes up factorially.

http://en.wikipedia.org/wiki/Factorial

That makes a 128 qubits equivalent to 385620482362580421735677065923463640617493109590223590278828403276373402575165543560686168588507361534030051833058916347592172932262498857766114955245039357760034644709279247692495585280000000000000000000000000000000 ordinary bits.

 

[wuliheron]

128 qbits? My knowledge of quantum mechanics is lacking, but shouldn't be possible to instantly break an equally long symmetrical key with a computer like that? E.G. AES-128 just became useless.

Exactly. These computers can't do every type of calculation faster then classical computers, but encryption is one place they excel.

The more interesting issue for me personally is the kind of research into fundamental physics they permit. One recent experiment with a much more modest quantum computer demonstrated that quantum mechanics is contextual, thereby ruling out a number of theories. A full fledged quantum computer like this is every theoretical physicists wet dream. We're talking major advances similar to all the technology that emerged from all the research done in WW II.

 

[dullard]

These are qubits, not bits. Like I said, their processing power goes up factorially.

I may be wrong, but from what I've read on it over the years, D-Wave doesn't actually have true qubits. They are close, but not quite.

Plus, just because I have four 16-bit computers next to each other doesn't mean they are as good as a single 64-bit computer even if they ran at the same frequency. The communication overload slows it way, way down. That appears to be what D-Wave has: 128 single qubit computers next to each other. Many people have speculated that it'll take over a 10000 "D-Wave almost-qubit" computer to equal a computer with today's PC power. They aren't anywhere near that.

If I'm out of date or wrong, please send me independent links showing it doing better than a standard computer on anything (even things that quantum processors would be perfect for). I'd like to read up more on it.

 

[tommo123]

why is it that quantum computers are good at tasks like encryption but not others in a general cpu sense? is it likely that these would be in a sense, an addon board to PCs to accelerate certain tasks only?

 

[wuliheron]

I may be wrong, but from what I've read on it over the years, D-Wave doesn't actually have true qubits. They are close, but not quite.

Plus, just because I have four 16-bit computers next to each other doesn't mean they are as good as a single 64-bit computer even if they ran at the same frequency. The communication overload slows it way, way down. That appears to be what D-Wave has: 128 single qubit computers next to each other. Many people have speculated that it'll take over a 10000 "D-Wave almost-qubit" computer to equal a computer with today's PC power. They aren't anywhere near that.

If I'm out of date or wrong, please send me independent links showing it doing better than a standard computer on anything (even things that quantum processors would be perfect for). I'd like to read up more on it.

I checked up on that and it seems you are correct. It apparently uses entangled pairs of quanta rather then entangling all 128 together, so it isn't a full fledged quantum computer.

 

[halik]

http://physicsworld.com/blog/2011/05/d-wave_sells_its_first_quantum.html

While laboratories around the world struggle to entangle a few quanta and use them for calculations these people have already managed a working 128 qubit system. To put this in context, a 30 qubit computer would equal the power of a 10 teraflop conventional computer and the power of quantum computing systems goes up factorially with each new qubit added. That's enough power to likely break any conventional encryption systems used worldwide including I suspect that of Wikileaks and most governments.

Europe has been investing in quantum cryptography for quite some time now, especially in the banking sector, which is the only type of cryptographic system theoretically immune to a quantum computer. For its part the US has discouraged even the use of some conventional cryptographic systems in the private sector. The obvious implication is that D-Wave's new computer is the first public shot in a new arms race which will likely last for many decades to come.

Thank god my 256 AES is safe

 

[halik]

128 qbits? My knowledge of quantum mechanics is lacking, but shouldn't be possible to instantly break an equally long symmetrical key with a computer like that? E.G. AES-128 just became useless.

It will enumerate the entire keyspace of AES-128 instantly.

 

[halik]

$10,000,000 for a computer that is outperformed by todays PCs on the one and only one task that it can do? I'll pass for now. Maybe give them 10 more years though and I may change my tune.

WAT?

Edit: Looks like you're right, they have pairs 64 2 qbit computer in series it looks like. Shouldn't that amount to the same thing though, since each of the 2qbit states is uncorrelated? ie it will still enumerate a full 128 bit sequence.

 

[Kirby]

yeah, but will it play crysis at max settings?

 

[totalnoob]

yeah, but will it play crysis at max settings?

Are you kidding? It will make Crysis do this. http://www.youtube.com/watch?v=MtcoKpz45Kg

 

[silverpig]

Back when I was doing physics I was in a lab in Vancouver full of people who actually understood this stuff. My understanding is that DWave is more hype than anything, but they do have a lot of money and a lot of people working on it full time, so who knows.

The last demonstration I saw several years ago made it look like they had a really fancy analog computer that could do some quantum stuff, but wasn't a full quantum computer.

I've met Geordie Rose a few times, as well as a few of the early investors. Geordie is a good guy and I hope it all works out for them.

 

[wuliheron]

why is it that quantum computers are good at tasks like encryption but not others in a general cpu sense? is it likely that these would be in a sense, an addon board to PCs to accelerate certain tasks only?

As I understand it quantum computers can process data much faster then a classical computer, but some types of calculations take longer to feed into the computer then they would for a classical computer. Like anything in life it seems there are always tradeoffs.

Addon boards for a cpu might be great for specific tasks such as decryption, but probably by the time it is possible to make one cheap enough and small enough for a PC it will either be integrated into cpu or merely added by plugging another chip onto the motherboard. The general trend is for the entire motherboard to migrate onto the cpu and for the cpu to incorporate more specialized circuitry. Sort of like the human brain where instead of just a calculator you have all sorts of specialized circuitry and most of the number crunching all bundled together to make it faster.

 

[wuliheron]

Thank god my 256 AES is safe

That's assuming the government doesn't have a secret program already going.

 

[wuliheron]

yeah, but will it play crysis at max settings?

Theoretically it will play Crysis at max settings in multiple universes simultaneously.

 

[wuliheron]

Back when I was doing physics I was in a lab in Vancouver full of people who actually understood this stuff. My understanding is that DWave is more hype than anything, but they do have a lot of money and a lot of people working on it full time, so who knows.

The last demonstration I saw several years ago made it look like they had a really fancy analog computer that could do some quantum stuff, but wasn't a full quantum computer.

I've met Geordie Rose a few times, as well as a few of the early investors. Geordie is a good guy and I hope it all works out for them.

It really does make you wonder if even a partial quantum computer is already commercially feasible just what is the government doing in secret?

 

[shira]

Crytographic encryption systems based on integer factorization would be vulnerable to quantum computers, but other encryption systems would not be.

For example, lattice-based cryptographic systems use a difficult mathematical lattice problem - the closest-vector problem is a common one - as the basis for the encryption. Quantum computational techniques are not effective at solving such problems. Presumably, then, long before quantum computers are commonplace, all high-security systems will have moved to encryption systems that resist quantum computations.

 

[Double Trouble]

Crytographic encryption systems based on integer factorization would be vulnerable to quantum computers, but other encryption systems would not be.

For example, lattice-based cryptographic systems use a difficult mathematical lattice problem - the closest-vector problem is a common one - as the basis for the encryption. Quantum computational techniques are not effective at solving such problems. Presumably, then, long before quantum computers are commonplace, all high-security systems will have moved to encryption systems that resist quantum computations.

Interesting observation, that makes sense. That is, of course, with the exception of quantum decryption systems that are likely to already be in use in some form at the NSA.

It's fascinating stuff, the quantum computing mechanisms boggle the mind (at least mine).

 

[Atreus21]

I guess there's no simple explanation for how the hell this stuff works, huh?

 

[Jaskalas]

I guess there's no simple explanation for how the hell this stuff works, huh?

A wizard did it.

BTW, anyone think we have interfaces capable of utilizing the data input and output for this sort of thing? I think the bus / data transfer / storage would be a catastrophic limitation.

 

[silverpig]

I guess there's no simple explanation for how the hell this stuff works, huh?

ArsTechnica had a really great article about it a few years back. I'm sure it's still up.

 

[Kirby]

A wizard did it.

BTW, anyone think we have interfaces capable of utilizing the data input and output for this sort of thing? I think the bus / data transfer / storage would be a catastrophic limitation.

"We spent all our money on this whizbang processor, so the only storage we can afford are 8" floppy disks."