IBM Plans to Commercialize 58-Qubit Quantum Computer
Norishige Morimoto, Director of IBM Research in Tokyo and global vice president at IBM, said that IBM intends to commercialize quantum computers within 3-5 years, when he expects quantum computers to outperform supercomputers in specific domains.
[...] The company's latest System Q One quantum computing system has a 20-qubit quantum processor with a quantum volume of 16. Quantum volume is a quantum computing performance metric IBM believes is more accurate than just using qubits alone. Quantum volume uses a combination of the number of qubits and error rate to determine the real-world performance of a quantum processor. The company is currently giving others free and paid access to its existing quantum computers.
IBM, Google and others have said before that to achieve quantum supremacy, a quantum computer needs at least 50 qubits. Morimoto said that IBM plans to launch a next-generation 58-qubit quantum computer that can outperform supercomputers and thus are suitable for commercialization.
However, don't expect to own one of these any time soon, as they will require a working environment with a temperature of -273 degrees Celsius to protect the qubits from interference. As such, IBM believes that this sort of quantum computer will work best as a companion to classical supercomputers.
Room temperature or bust.
Previously: IBM Announces Working Prototype of a 50-Qubit Quantum Computer
IBM Announces its First Commercially Available Quantum Computer (20 qubits)
IBM's New Quantum Computer Will Have You Drooling
(Score: 2) by jmorris on Tuesday May 28 2019, @09:38PM (6 children)
> Room temperature or bust.
Nah, if it needs to be cooled it is still commercializable.. What becomes a problem is when it needs these sort of extremely low temps that require regular delivery of liquefied gases or an onsite production facility. Gets kinda spendy either way, even in the rarefied budgeting associated with high end computing.
I'm questioning the 50 qubits is "quantum supremacy" assertion. What can actually be done with only 50 bits? You couldn't even try to crack an AES key, could you? And no public key system ever used keys that short. Suspect modeling nukes also needs a few bits as well. Thinking this is just another "We gotta beat the Reds to this" WE MUST NOT LOSE THE RACE FOR QUANTUM SUPREMACY!! IT WOULD BE LIKE LETTING IVAN LAND ON THE MOON FIRST!! GIVE US SACKS OF CASH OR YOU DON'T CARE IF AMERICA LOSES TO CHINA!!
(Score: 2) by isostatic on Tuesday May 28 2019, @10:14PM (2 children)
Quite.y ice cream doesn’t work at room temperature, yet it’s commercialised.
(Score: 2) by krishnoid on Tuesday May 28 2019, @11:20PM
In fact, it's even creamier when it's liquid-nitrogen cooled ... wait a second ... quantum supercomputers that also make extra-creamy ice cream? Quite.
(Score: 2) by stormwyrm on Wednesday May 29 2019, @04:32AM
Numquam ponenda est pluralitas sine necessitate.
(Score: 2) by takyon on Tuesday May 28 2019, @11:12PM
I want 1 million qubit chips in smartphones with no cooling.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 0) by Anonymous Coward on Wednesday May 29 2019, @12:15AM (1 child)
A bit and a qubit are very different things. 50 qubits would have factoral 50 computational outcomes. You could easily break 4k RSA keys with that. Assuming of course this is a general purpose quantum computer and not yet another annealer
(Score: 4, Informative) by stormwyrm on Wednesday May 29 2019, @03:59AM
No. 58 qubits is nowhere near enough to break any RSA keys that were ever in use. To use Shor's algorithm to factor a number you have to have at least twice the number of qubits as there are bits in the key, probably a little bit more because you want some error correction too. With only 58 qubits you might be able to factor a number of about 29 bits or so, maybe a number like 394,245,403 you'd be able to factor that way. Much easier and faster to do trial division on a normal computer to factor a number like that. A long, long way from the kinds of moduli that were ever used for RSA in the real world.
It's far from useless however. A 58-qubit quantum computer however would probably be very good at molecular dynamics simulations and such though: such quantum-mechanical systems are notoriously expensive to simulate on classical computers even when the number of particles involved is still fairly modest. It might even be able to simulate certain kinds of systems involving the other fundamental interactions that are difficult even for the most powerful classical computers of today.
Numquam ponenda est pluralitas sine necessitate.