Engineers Just Unveiled The First-Ever Design of a Complete Quantum Computer Chip

a reply to: SprocketUK

Ultraviolet blue screen of death?!!!



As long as there is no Y2K bug built in!

QC will have to integrate with regular computers. If people think AI programmers are scarce then the QC programmer will be more rare.

My brief reading in the topic is that this is a different kind of thinking. You do not code to solve a problem but you look at data and explain the solution. Computer sciencists don’t think that way. The field is going to be so specialized it is not even funny.

originally posted by: TEOTWAWKIAIFF
a reply to: SprocketUK

Ultraviolet blue screen of death?!!!

As long as there is no Y2K bug built in!

QC will have to integrate with regular computers. If people think AI programmers are scarce then the QC programmer will be more rare.

My brief reading in the topic is that this is a different kind of thinking. You do not code to solve a problem but you look at data and explain the solution. Computer sciencists don’t think that way. The field is going to be so specialized it is not even funny.

the fact that you don't understand something, doesn't make it an arcane art for everyone else. it doesn't give you a free pass to spread BS either.

How to program a quantum computer
Quantum programming languages

a reply to: TEOTWAWKIAIFF

Once you get it right, AI that is, every new iteration of the thing will be designed by itself.

Even the people who manage to design such will probably not be able to comprehend, in its entirety, how future version of said code function down to the nature, complexity and speed of every new iteration.

Programming quantum systems may be rather different but one way to try and understand the quantum bit(Qbit) is to consider what would happen if we could split a function like boolean negation into 2 equal parts.

a reply to: andy06shake

the problem with AI is that you don't have to understand how it works (and with the complexity of some networks, it's sometimes hard to tell), you only need to know how to get there, by training the network. that's of course a huge simplification, because there's a matter of input/output data, transformations and interpretations of said data and so on, but the core remains.

with AI developing new methods of training the network, or designing the network itself, even the "how did we get there" part goes out of the window.

a reply to: jedi_hamster

It's beyond my ken anyway.

Back when i learned to code it was all Pascal and C++.

These days i can read the stuff and perform basic edits/modifications but that's about it. LoL

Still, i imagine some form of logic that fits will materialize to facilitate the construction of Quantum applications.

a reply to: TEOTWAWKIAIFF

There will only every be one generation of programmers that get AI right. After the first time, as others say, it will start programming more itself.

As for the rest of this thread, I remember the talk of regular computers being the size of soccor field and now we are talking about that with quantum computers. Wait a few years and we will have them the size of everything else these days.

a reply to: jedi_hamster

Challenge Accepted!

A quantum state is the information needed to completely describe a particle. Since the design of OP is electron spin in silicon I’ll stick to those.

Electrons can exist with any spin orientation. That means 3D degrees of freedom. Called a steradian or “solid degrees” that would be necessary for a 3D object but quantum measurements are not that complicated.

Polarization has three states. Horizontal or vertical polarization; diagonal or not diagonal polarization; then right and left circulation. Those are the measurements you get without having to deal with the entire notion of 3D’edness. There is a guy in Russia who has the idea of using the 3D’edness allowing for more states to be observed in a qubit (5 states, IIRC). But keeping it simple makes the math easier.

Spin, in and of itself, is even easier with two choices, up or down.

Armed with these notions you make measurements at the quantum level to use your electron qubit. But then it gets even better because you can use the quantum effects between electrons for your computing needs. Called quantum entanglement your qubit pairs up with a buddy which has the opposite state. So, if we stick with spin, when one is spinning up, the other is spinning down.

Each measurement changes the state the electron is in. Knowing you have entanglement allows you to infer the state of the other qubit. You can entangle as many electrons together as you are capable of doing.

The math used to describe these arrangements is known as linear algebra. Because operations on vectors are in matrix form the math is different than regular, every day algebra. Matrices are non-communicative meaning the order in which you perform you operations is not swappable. AxB is not the same as BxA. You are also using real numbers (i.e., complex math, square root of -1). Each operation changes states of your entangled electrons. You let the system settle to a state then read your qubits which is your answer.

You need to know physics and algorithm construction (computer science) to set your quantum computer up to make meaningful statements. This is not
10 PRINT “Hello world!”

This programming is far different conceptually. Allowing multiple qubits to “run” to a state is wave-particle duality interference. Electrons do not exist. Weird huh? But that is what quantum mechanics tells us. Electrons exist as a possibility in a complex equation. It can possibly not exist as well. You do not know until you look.

Still think I’m full of it? Still think I don’t know what I am talking about? Go look for yourself. Does that “alpa|0> + beta|1>” Wikipedia make sense to you? Quasi C++ shows how set a qubit. I'm talking about the algorithm itself. I’ve been reading it now for months and still need to pause and remember what is going on. Good job on calling me out on my belief system! Next time, follow the teaching of Lao Tzu before rushing into battle!

I've got a degree in computer science, a minor in math. Particle physics is something I just grok. Had I had access to money it may have been what I actually studied in college instead being practical and getting a degree to get work.

Gizmodo.com - What the Hell Is a Quantum Computer and How Excited Should I Be?. (see, "These equations can look kind of scary, like this" for an example of a complex matrix operation)

a reply to: TEOTWAWKIAIFF

I will give you the benefit of the doubt and assume that your title is not your doing.

1) It is NOT the first ever design of a quantum computer - there are actually several complete designs from other groups approaching the problem from different angles. There are even several that have been built - IBM has a 50 qubit example in operation. A practical machine will need several million qubits - and there is the rub: how to scale it up in a practical matter. UNSW claims its design will be easily scalable.

2) It is ONLY a design - it has NOT been proved. As I said, IBM, and others, have a working demonstration of its design in operation and the race to scale it up is well and truly under way. UNSW does not yet have even a demo, and it is still a long way off.

3) What this is is a THEORETICAL plan to build QC's on silicon, something that we are very familiar with. If it works, and it is indeed scalable, then bingo. But we ain't there yet.

For more info read here: UN SW claims first complete silicon quantum computer chips, others not so sure

Here is the money quote from that article:

Silicon is the material used to build our existing computer chips; this chip could theoretically be manufactured using existing equipment.

Theoretically is the key word, however.

"This is not a demonstration of the design, it's just a design. It has not been proven. There are plenty of designs for quantum computers. Everyone believes they have a scalable architecture," says Christopher Ferrie​, a researcher at the Centre for Quantum Software and Information, the University of Technology Sydney's competing quantum venture.

Scott Aaronson​, a top international quantum computing researcher at the University of Texas, puts it even more bluntly.

"It's sort of a running joke in the field that scalable quantum computers were already built more than a decade ago – at the level of impressive-looking PowerPoint animations.

"The challenge is to make it real."

originally posted by: stormcell

originally posted by: chr0naut

originally posted by: TEOTWAWKIAIFF
a reply to: chr0naut

They can do quantum encryption now using D-Wave.

...

Yes, but what happens to the value of all conventionally mined coin when they start mining with quantum computers?

The entire theoretical maximum coins would be generated so quickly that conventional coin would tumble.

It would go the other way first . As the coins were mined, the price would go up and up, everyone will keep holding on to their coins as the prices go up. They might diversify into other cryptocurrencies to spread their bets.

With the way quantum computers work bitcoins become useless. A quantum computer would not unlock a single block, it would unlock them all at once. You get to try every probability of the equation at once meaning it would take secondsto shake out the right answers.

originally posted by: TEOTWAWKIAIFF
Still think I’m full of it? Still think I don’t know what I am talking about?

looks like it.

see, you've missed one thing. one thing that was actually mentioned in one of the links i've posted. quantum computers are - and will be - hybrid devices.

which basically means you're programming them just like any other modern cpu-based device, with the added bonus of having a "quantum coprocessor" available for any calculations you wish to run on it. how many calculations will you run on it - that's up to you, and it's obvious that programming it requires different approach, but it isn't exactly rocket science either. basic math and logical thinking are enough to get started.

in general, the code isn't some exotic mumbo-jumbo like you're trying to portrait it - it's a mixture of current code with code using quantum features, which is why they're even adding quantum features to existing programming languages that were around decades ago. any good programmer that understands the underlying concepts should be able to handle it.

so yeah, you're full of it. first you don't read the paper you're posting, then you're trying to build up hype like the design is something verified, 100% perfect and ready for production - while it's only theoretical and requires technological advances to be even verified in practice - and in the end, you're switching to FUD, claiming that programming a quantum computer is some arcane knowledge reserved to few.

originally posted by: TEOTWAWKIAIFF
You are also using real numbers (i.e., complex math, square root of -1).

that alone says you have no clue what you're talking about. real numbers are real numbers, complex numbers are complex numbers. square root of -1 is an imaginary number.

originally posted by: TEOTWAWKIAIFF
This is not
10 PRINT “Hello world!”

of course it is - if you want it to be. you can use any regular programming language that has added support for quantum computing, with all its features.

sure, the real magic begins when you start using qubits to do the dirty work - as i've said, developing new algorithms using qubits requires different approach, but it isn't beyond a good programmer worth his salt. developers from all over the world were running over 200 thousands of experiments during last month alone using IBM's Q Experience program - clearly, they're all idiots and have no idea what they're doing, right?

quantumexperience.ng.bluemix.net...

What if a quantum computer gets into the mind of humans and makes them build a few robots , which in turn builds the next nuclear arsenal and destroys the planet ?

Very possible!

originally posted by: Denoli
What if a quantum computer gets into the mind of humans and makes them build a few robots , which in turn builds the next nuclear arsenal and destroys the planet ?

Very possible!

...NOT.

what's more possible is that talking out of your rear will make you look like a fool.

originally posted by: jedi_hamster

originally posted by: Denoli
What if a quantum computer gets into the mind of humans and makes them build a few robots , which in turn builds the next nuclear arsenal and destroys the planet ?

Very possible!

...NOT.

what's more possible is that talking out of your rear will make you look like a fool.

Sorry jedi hamster pffft

Update at UNSW

"In placing our phosphorus atoms in the silicon to make a qubit, we have demonstrated that we can use a scanning probe to directly measure the atom's wave function, which tells us its exact physical location in the chip. We are the only group in the world who can actually see where our qubits are.

"Our competitive advantage is that we can put our high-quality qubit where we want it in the chip, see what we've made, and then measure how it behaves. We can add another qubit nearby and see how the two wave functions interact. And then we can start to generate replicas of the devices we have created," she [Simmons] says.

Study lead co-author, UNSW's Sam Gorman, says: "Theory had predicted the two qubits would need to be placed 20 nanometres apart to see this correlation effect. But we found it occurs at only 16 nanometres apart.

"In our quantum world, this is a very big difference," he says. "It is also brilliant, as an experimentalist, to be challenging the theory."

"Our results confirm that silicon is an optimal choice, because its use avoids the problem most other devices face of having a mix of different materials, including dielectrics and surface metals, that can be the source of, and amplify, electrical noise.

"With our precision approach we've achieved what we believe is the lowest electrical noise level possible for an electronic nano-device in silicon - three orders of magnitude lower than even using carbon nanotubes," she says.

Phys.org, March 7, 2018 - Seeing is believing—precision atom qubits achieve major quantum computing milestone.

First, @OP, it was their title. Yes, there have been designs for quantum computers for a while. The first I was aware of was the size of a soccer field. But the headline is for a complete design "in silicon". They are etching the circuitry needed for the non-quantum portion into currently available manufacturing processes. That has not really been attempted as a universal quantum computer.

Second, @JediHamster, OK. But so what? I don't have to be a master of something to bring it up for discussion. And I still stand by my statement that pseudo code is not really quantum computing. Render out Shore's algorithm in linear algebra and maybe I will believe you.

Anyway, back to UNSW and the news. I wonder what they are feeding the researchers?!!

First, the main researcher, Michelle Simmons, was named Australia's Person of the Year earlier this year. The tolerance at where they can keep these qubits (phosphorus atom, not the spin qubits of OP) is amazing. Knowing exactly where they are and being able to see the wave function of the atom is BA! They are using the same technique as before (etching circuitry into silicon) but using quantum dots and phosphorus atoms at a distance of 16 nanometers. 4 nanometers difference may not seem like much but being better than theory is!

Second, they are building this at the atom level! They are not only building it atom by atom, they can see and measure it as well. Pushing things around at the atom level is just fascinating. Watching correlation happen... what would cooler than that?

Doing both the electron spin and the quantum dot qubits at the same university is very d@mn impressive!! Both are being pursued in parallel and have patents pending for each.

a reply to: TEOTWAWKIAIFF

Pretty cool stuff.

Given the exotic nature of it, though, I would expect a high retail price if ever it comes unto the market.

I wonder if it'd be possible to mimic the ability of quantum computing using a new programming language...

a reply to: swanne


Luckily it is a university. They have government funding and co-research facilities around the globe.

They have a spin out company too. They are shooting for 2022 for a 10-qubit demo model. But the really cool part is doing much of the chip design in silicon. Looking around at all the cell phones gives an indication how significant doing this in a proven technology really is.

As far as mimicking quantum computing goes... they have real world physics simulators up to 50-qubits. I've seen a web based game that is actually doing quantum calculations. So yeah, there probably is a way to do it in a programming language. The question then becomes how long would it take to make, take to run, and how much storage would be needed.

Why re-invent the wheel? Rigetti Computing has web-based access to their 14-qubit system using a programming language called Forest. This April, they will demo their 19 qubit computer. They also have a virtual one available for use now via the cloud.

IBM has their Q-perience also available. Google claims to have a 72-qubit chip (you would have to purpose some for error correcting but they are hoping you would end up with ~ 40 qubits left over, or more).

Add the UNSW news and the future of quantum computing is getting closer, faster than some realize!

a reply to: TEOTWAWKIAIFF

Hmmm, I should check out those, thanks for mentioning these.

I know a good deal of JavaScript, so in fact I am thinking about trying something myself.

a reply to: chr0naut
Not really. The release of new Bitcoin is based on a calander timeframe. More CPU power just means you take a bigger piece of the pie when they are mined. More computing power wouldn't speed up the rate that Bitcoin are added to circulation.

Progress in this field is so rapid it's hard to keep up. Google just unveiled their 72 quantum bit Bristlecone - not a blueprint but a working processor. They plan to test it over the coming months in an effort to prove it can solve problems no classical computer can.

The engineering challenges are so huge there is much cooperation with incremental breakthroughs discussed and published. The diverse approaches make it all the more interesting.

Other research reported today:
Topological superconductor phase may solve decoherence problem in quantum computers

the phase, once integrated into a suitable material, would be capable of supporting Majorana bound states

This may have implications for the topological fermion-based computer Microsoft are placing their bets on.

And...

Scaling silicon photonic quantum technology

...the team has found a way to generate and entangle pairs of particles that each has 15 states.

and this with 'unprecedented precision' on a single silicon chip, paving the way for photonic quantum computing.

Also yesterday:
How a string of giant atoms might bring quantum computers to physics labs
Discussing the development of analogue quantum computers using trapped Rydberg atoms.