IBM Publishes its Quantum Roadmap, Says it Will Have a 1,000-qubit Machine in 2023 (techcrunch.com) 35
IBM today, for the first time, published its road map for the future of its quantum computing hardware. There is a lot to digest here, but the most important news in the short term is that the company believes it is on its way to building a quantum processor with more than 1,000 qubits -- and somewhere between 10 and 50 logical qubits -- by the end of 2023. From a report: Currently, the company's quantum processors top out at 65 qubits. It plans to launch a 127-qubit processor next year and a 433-qubit machine in 2022. To get to this point, IBM is also building a completely new dilution refrigerator to house these larger chips, as well as the technology to connect multiple of these units to build a system akin to today's multi-core architectures in classical chips. IBM's Dario Gil tells me that the company made a deliberate choice in announcing this road map and he likened it to the birth of the semiconductor industry.
"If you look at the difference of what it takes to build an industry as opposed to doing a project or doing scientific experiments and moving a field forward, we have had a philosophy that what we needed to do is to build a team that did three things well, in terms of cultures that have to come together. And that was a culture of science, a culture of the road map, and a culture of agile," Gil said. He argues that to reach the ultimate goal of the quantum industry, that is, to build a large-scale, fault-tolerant quantum computer, the company could've taken two different paths. The first would be more like the Apollo program, where everybody comes together, works on a problem for a decade and then all the different pieces come together for this one breakthrough moment.
"If you look at the difference of what it takes to build an industry as opposed to doing a project or doing scientific experiments and moving a field forward, we have had a philosophy that what we needed to do is to build a team that did three things well, in terms of cultures that have to come together. And that was a culture of science, a culture of the road map, and a culture of agile," Gil said. He argues that to reach the ultimate goal of the quantum industry, that is, to build a large-scale, fault-tolerant quantum computer, the company could've taken two different paths. The first would be more like the Apollo program, where everybody comes together, works on a problem for a decade and then all the different pieces come together for this one breakthrough moment.
Definitions (Score:5, Informative)
Quick refresher - a qubit is, roughly, the quantum equivalent of a bit. A logical qubit is the quantum equivalent of a logic gate. The way it was explained to me by a physicist is a qubit is similar to a diode that can store one bit of information, and you group them together, as you would a regular diode, into transistors, then group those together to make a logic gate. Not a 1==1 analogy, but close. So IBM's machine would have 50 quantum-equivalent logic gates. Maybe enough for a very simple accumulator and bit shifter, or something similar.
640 qubits. (Score:1)
640 qubits ought to be enough for anybody.
Re: (Score:2)
Aehm, diodes cannot store information. You either need capacitors (inductors could work but are impractical) or full-fledged Flip-Flops involving two transistors, or, lately, phase-change materials.
Re: (Score:2)
similar to a diode that can store one bit of information
diodes cannot store information
But quantum computing is about as real as a diode storing information, so the comparison is logical.
:D
But a diode can store data, overload it and burn it out and now it is in another state. Write once but it can be read many
Re: (Score:3)
Quantum computers with thousands of qbits can potentially do things quickly that classical computers with billions of bits are very slow at. It is a good start toward an entirely new technology. Quantum computers won't replace classical computers, but they are likely to do some revolutionary things in
Re: (Score:2)
And an important data point: quantum computers don't even have memory; there is no data storage at all. The machine runs, and now has to be reset. Any work must be output during that run, or it was lost.
Re: (Score:2)
Yes, fair bit of confusion here. Diodes can form logic gates (https://en.wikipedia.org/wiki/Diode_logic) but can't store information. (except the nice burn it out diode storage in another comment :) ).
Tunnel diodes and 4-layer diodes have been used as memory elements in special applications.
Re: (Score:2)
Wrong. Tunnel diodes in an appropriate circuit have two stable states, and are thus capable of storing one bit. Alas, such a memory cell is very inefficient.
Re: (Score:2)
Re: (Score:2)
Wrong. Tunnel diodes in an appropriate circuit have two stable states, and are thus capable of storing one bit. Alas, such a memory cell is very inefficient.
But they are very fast and very radiation resistant.
Re: (Score:2)
Just out of curiosity, how does one pronounce "Qubit"? Is it "kwibit"?
I read it as "cubit." But then I think of the old Bill Cosby bit about Noah [youtube.com] and I'm not so sure...
Re: (Score:2)
This is incorrect. A logical qubit is an entangled collection of physical qubits that you use to do calculations as though it is a single qubit. The point is that you can use the multiple physical qubits that make up the logical qubit to do error correction.
Quantum error correction is pretty much the biggest topic in quantum computing right now, because the science for making the hardware isn't there yet, so error rates are much too high to do anything practical.
Look for quantum computers at the checkout aisle (Score:1)
There won't be anything other than government-ran quantum computers for another ten years. That's 2030.
There won't be consumer quantum computers for another five years after that. That's 2035.
There won't be "quantum graphing calculators" in your grocery checkout aisle ever.
Quantum computing has its uses, and *none* are consumer-level. The lowest level you'd find are a replacement for GPUs, and high-speed determinations and derivations. None of these will aid cars, mobile phones, tablets, or home PCs.
It
World market. (Score:4, Funny)
There is no reason anyone would want a quantum computer in their home. I think there is a world market for maybe five of these computers. Regarding processing power, 640 qubits ought be enough for anybody.
Re: (Score:2)
Cute.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
do we still make computers the same way as the ENIAC, or perhaps has time moved on?
I mean, I suppose you'd be arguing that we just use really tiny vacuum tubes instead of semiconductor devices...
Re: (Score:2)
Your argument is reasonable...but wrong.
OTOH, as long as quantum computers need cryogenic storage you won't find people among the customers. You can't associate that with a date.
FWIW, we don't know the capabilities of quantum computers because we've never played around with them. All we know they can do well are a few edge cases, like factoring numbers and molecular modeling, but that's no argument that that's all that they can do better, or even do well. It could be true, but there's no reason to believ
Re: Look for quantum computers at the checkout ais (Score:2)
There are consumer Quantum computers now... All IBMs Quantum computers are accessible via the cloud. All the future ones will be as well.
The entire business model of Quantum is going to be driven by cloud. It has a lot of parallels to early mainframes.
Re: (Score:2)
Re: (Score:3)
The former is noisy and decoheres randomly, and needs an error correction logical layer on top of it to get "logical qubits" with which one can do computation. Those estimates are for logical qubits, and given IBM predicts just 10-50 logical to the 1000 physical, then elliptic curves are safe for at least 8 more years than you claim, if Moore's Law applies.
Re: (Score:2)
if Moore's Law applies.
That is exceptionally unlikely. Remember, they need to keep entangling going. Moore's law is for computing elements that are only coupled but wires and do not share state of any kind.
Re: (Score:1)
Today is the first day I've heard people differentiate physical and logical qubits. The only reason I can infer why IBM does this is because 1000 "qubits" makes for better PR than 10 actual qubits. It strikes me as misleading because when people think in qubits they are thinking in *actual quantum bits* in a universal quantum logic circuit, not whatever niche exotic quantum calculator google or IBM built that can't actually do arbitrary quantum logic operations.
When you read the amazing journal reports that
Re: (Score:2)
1000 "qubits" makes for better PR than 10 actual qubits
Maybe we should rename the logical groupings as "quibibytes" but don't tell anyone. Then use "qulobytes" on the product packaging so it looks like more.
PS: products aimed at those finicky qbit-o-philes can point out how their purified entanglements are proven to be smoother and less tinny than the leading brand.
Re: (Score:2)
for RSA4096 theres still a way to go to crack it, assuming IBM is serious about this 1000 qubit machine. However, ED25519 and elliptic curves only requite 1500 or so qubits so if this system ever comes to fruition in 2023, the elliptic curve ship has sailed and we'll need something better. post quantum crypto is so far in the hands of Microsoft and Google, so i suspect Dan Bernstein or someone of equal capability will need to rise to the occasion to deliver something the NSA hasnt stuck their thumb in first.
Nope. Qbits and qbits that actually stay entangled under complex calculations are two pretty different things. They are nowhere near where you think they are.
Re: calls for replacements to ED25519. (Score:2)
Not just Google and Microsoft
https://www.zurich.ibm.com/sec... [ibm.com]
Road map (Score:1)
I'm not sure I trust their quantum-computing chops when they've clearly failed to take both paths simultaneously.
Re: (Score:2)
+1
Don't read too much onto this. (Score:2)
A 1000 qbit machine with today's technology is one thing.
A 1000 quantum error correction qbit machine on which you can perform logic is a completely different thing.
The latter doesn't exist and may never exist. There is reasonable skepticism that quantum computers will never do those things. It has not yet been settled.
The 'quantum supremacy' thing a few months ago was on a problem small enough that a logic-capable quantum error corrected qbit computer was not needed. It proved a principle but did show prac
Quantum computing is cool... but Quantum.. (Score:2)
..communication would be a game changer.
It'd be great to get our private communications back.
A long way to go. (Score:2)
so what? (Score:1)