IBM shows off quantum computing advances, says practical qubit computers are close

[Anonymous No.4500163]

>IBM shows off quantum computing advances, says practical qubit computers are close

>Quantum scientists at IBM Research have announced major advances in quantum computing that could place real, practical quantum computers in businesses and homes within the next 10 years.

>The main breakthrough revolves around the long-term integrity of qubits. To perform quantum computing, you need to be able to reliably store and interrogate qubits — but qubits are incredibly flighty creatures that readily change their state through decoherence. IBM has created a high-coherence 3D qubit that retains its state for up to 100 microseconds, or 0.1 milliseconds. This is stable enough that engineers can now shift their focus to scaling up the number of qubits to create a quantum logic computer.

>In a separate experiment, IBM created a 3-qubit chip using conventional semiconductor fabrication techniques, which was then used to perform a controlled-NOT logic operation with a 95% success rate. This kind of success rate is unheard of in quantum computers, and again comes down to the fact that IBM has managed to create qubits with relatively high stability (10 microseconds in this case). IBM is confident that this experiment paves the way to multi-qubit logic chips.

[Anonymous No.4500167]

>>4500163
Call me when there's a C compiler

[Anonymous No.4500169]

>All in all, IBM Research is now saying that bona fide quantum computers are now just 10 to 15 years away. Why is this significant? Well, put it this way: According to IBM, 250 qubits would be able to store “more bits of information than there are atoms in the universe.” This in itself is truly awesome — but then when you factor in that a quantum computer could perform logic on all of that data, in parallel, instantaneously… well, you begin to see the power of quantum computing. You’re talking about the performance of a supercomputer on a single chip.

>But therein lies the crux: We’re still not entirely sure what we would use quantum computers for. It is generally believed that quantum computers would be the most beastly factorization machines ever conceived. Factoring is the basis for modern cryptography. Quantum computers would not only excel at secure communications, then, but they would also be able to break any non-quantum-encrypted data in seconds.

>Beyond that, quantum computers might be good at investigating databases of unstructured data (similar to the quantum PageRank), and attempting unsolved mathematical problems.

[Anonymous No.4500194]

Will they be useful for computer games?

[Anonymous No.4500199]

I wish I understood how qbits and shit work

How is a bit thats "both on and off" even useful?

[Anonymous No.4500209]

>that feel when Moore's Law holds after the death of silicon

[Anonymous No.4500226]

Didn't IBM also build a CPU out of graphene or something?

[Anonymous No.4500233]

>>4500163
But can it play Crysis?

[Anonymous No.4500242]

>>4500209
Moore's Law has, and will always, hold after the death of architectures and manufacturing processes. Every time someone has said "No, this will not continue since we are at the theoretical limits of these systems", something new always came along.

>that feel

[Anonymous No.4500248]

>>4500242
Well nothing as major as silicon hitting its limit has ever come

[Anonymous No.4500257]

>>4500169
> 250 qubits would be able to store “more bits of information than there are atoms in the universe.”
I'm pretty sure it's the other way around

[Anonymous No.4500262]

That feel when Moore's law only holds because processor companies make a concious effort to make it hold.
The feel when they could make faster processors if they wanted to but don't.
That feel when you worship Moore's law even though it slows progression of technology down

[Anonymous No.4500282]

>>4500262
care to elaborate?

[Anonymous No.4500295]

>>4500282

He is saying Moore's Law is self-fulfilling.

I doubt he has any evidence.

[Anonymous No.4500306]

The real question is not, can you make a better CNOT gate, but has your reliability reached the point where you can do quantum error correction? That's what will enable quantum computers to scale, and until then, they're pretty useless. I can't tell how close they are to that from the blurb, but hopefully this is a big step closer.

[Anonymous No.4500317]

>>4500199
I'd love to know myself, I've been trying to understand for a year or so now.

[Anonymous No.4500319]

>>4500262
Yes! A conspiracy! Moore's law is nothing more than a price fixing scheme. If they gave us all the goods at once, they couldn't convince us to buy a new computer every 5 years. They have to draw it out.

[Anonymous No.4500324]

>>4500319
No, it's more of a "we've hit our target improvement now so we can relax for a few months" kinda thing

[Anonymous No.4500328]

>>4500324
That's what they WANT you to think!

[Anonymous No.4500340]

>>4500306
As far as I know people were able to do some rudimentary error-correction (which, as it turned out, created more errors than corrected); so yes, apparently, gates fidelity is a most important problem right now.

[Anonymous No.4500358]

What exactly are quantum computers? It's definition evades me. I keep hearing "OH IT HAS ONE AND ZERO AT THE SAME TIME", but what the fuck does that mean and how does it work?

[Anonymous No.4500364]

>>4500262
>The feel when they could make faster processors if they wanted to but don't.
>implying processors are faster now than six years ago

Single-core wise an old processor still kicks the crap out of any of those baby toys they put in computers now. Fuck this multithreading bullshit.

[Anonymous No.4500363]

>>4500169
I was under the impression that quantum computing could only trivialize symmetric encryption. The time to brute force asymmetric encryption would only be cut in half which is significant but doesn't mean much if all it does is take the time down from "the universe will undergo heat death before you crack this encryption" to "the Sun will be dead before you crack this encryption".

[Anonymous No.4500374]

>>4500262
Yes, I'm sure AMD is only letting its ass get kicked by Intel just to keep up appearances. If they wanted to they could have a 2000 core 550THz CPU out tomorrow, they just don't want to because losing to Intel's sandy bridge processors is so much better for business.

[Anonymous No.4500380]

>>4500374
I'm pretty sure my 4GHz i7 slaughters the performance of my old 2.4GHz FX-53 even in single threaded applications.

[Anonymous No.4500382]

>>4500374
AMD does anything but steal Intel's shit now?

[Anonymous No.4500383]

>>4500364
Thats because the Ghrz are the same, the new paradigm are the number of cores.

[Anonymous No.4500388]

>>4500358
It has two bits and qubit.

On, off, and on+off.

Thats as much as I can tell you. Through some voodoo magic it can do alogrithms extremely fast compared to older computers.

For example, in path finding you basically need to use a method of searching every possible situation. While this could take hundreds of calculations, quantum computers can somehow find the fastest route in a fraction of the amount of computations

[Anonymous No.4500390]

>>4500233
You need a cosmic AC for that.

[Anonymous No.4500396]

>>4500382
>AMD does anything but steal Intel's shit now?
That's all AMD ever did.

AMD came into business when they cloned the 8086.

They tried suing AMD but you can't patent numbers, so from that day on Intel used the "Pentium" brand on its x86 processors

[Anonymous No.4500426]

>>4500340
Yes, but the question I'm asking is how close are they to being good enough?

[Anonymous No.4500433]

>>4500396
False.
I remember several years ago when AMD had CPUs with LESS energy consumption, less temperature and overall better performance.

Those were the days when AMD was just superior, ofc Intel owns about 90% of the market share thus easier dominating it and buying off AMDs technologies.

And the reason it has so large percentage is cause its older in desktop CPUs and market their CPUs succesfully with ready computers, they had better contracts with those companies.

[Anonymous No.4500437]

>>4500363
You have asymmetric and symmetric mixed up, and you should have a square root where you have half, but otherwise that's right.

[Anonymous No.4500474]

>>4500433

And it paid Dell not to offer PCs with AMD chips in them

[Anonymous No.4500494]

>>4500364
what is power consumption

[Anonymous No.4500498]

>>4500494

something poor people worry about.

[Anonymous No.4500626]

>>4500163

Would these new quantum computers change the way we program? Would our current programming languages become obsolete?

[Anonymous No.4500653]

>>4500626
In the same way that more than one core changes the way we program. Quantum computing is better than traditional computing in certain situations, but not all.

[Anonymous No.4500662]

how good would these be for highly mathematical simulations of say, galaxy mergers or superclusters?

[Anonymous No.4500679]

>>4500653

If I misinterpret your words, I offer an apology.

Does this mean we could use languages like C and C++ with quantum processors as long as there is a compiler for them? The way these new processors work with data won't alter the languages we use? I mean, perhaps in the worst case scenario, we'd have to change a few things (like it happened with multicore programming).

[Anonymous No.4500688]

>250 qubits would be able to store “more bits of information than there are atoms in the universe.
270 normal bits can already store more information than there are atoms in the universe.
i am not impressed.

[Anonymous No.4500788]

So, when will there be affordable consumer level quantum computers that we can replace our desktop/laptops with?

[Anonymous No.4500797]

>>4500679
Functional/logical languages like Prolog and LISP might scale easily, but for largely procedural languages, I think there will be changes. There are algorithms that can be performed on quantum machines which are not feasible on regular ones. Now, there may be ways to simply translate by parallelizing, but ultimately it will be more efficient just to use new sets of languages which are specifically designed to take advantage of quantum machines.

C++ is starting to disappear anyway. Just a matter of time.

[Anonymous No.4500806]

so does this technology mean my iPhone will be faster or does it mean that I can store 100,000 more songs on my iPod?

[Anonymous No.4500812]

>>4500806

In the beginning, it will mean faster processing speed for anything it's applied to. Eventually (probably never) it will mean perfectly secure phone calls and VOIP.

[Anonymous No.4500839]

>>4500788
It said in 10 years.

[Anonymous No.4500849]

>>4500688
Actually no they can store a higher number but this is not information. 270 bits can store 270 pieces of information.

[Anonymous No.4500854]

>>4500688
>270 normal bits can already store more information than there are atoms in the universe.

[Anonymous No.4500859]

Close-up for anyone interested

[Anonymous No.4500873]

>>4500859
ugly color, pointy pins, too many lines 2/10 would not compute on.

[Anonymous No.4500882]

>>According to IBM, 250 qubits would be able to store “more bits of information than there are atoms in the universe.”

What does this even mean? I can't think of any non-ridiculous interpretations where this is actually true. Care to enlighten me?

[Anonymous No.4500886]

>>4500882
>http://www-03.ibm.com/press/us/en/pressrelease/36901.wss
>The special properties of qubits will allow quantum computers to work on millions of computations at once, while desktop PCs can typically handle minimal simultaneous computations. For example, a single 250-qubit state contains more bits of information than there are atoms in the universe.

[Anonymous No.4500888]

What about D-wave's quantum computer?

Is that lies? Lockheed Martin has baught one..

[Anonymous No.4500892]

>>4500882
Because of superposition, 250 qubits can store 2^250 pieces of information. That is, such an arrangement can enumerate more atoms than there are in the universe. It cannot simulate them all of course (mass, position, velocity are all extra pieces of info).

[Anonymous No.4500893]

>>4500888
It's a completely different type of machine. You wouldn't be able to run Shor's algorithm on it. Not that you could run Shor's algorithm on IBM's CNOT gate, but there it's just a matter of scaling things up successfully.

[Anonymous No.4500898]

>practical quantum computers in businesses and homes within the next 10 years.

no.

[Anonymous No.4500904]

>>4500892
>wolfram 2^250
>number name
1 quattuorvigintillion, 809 trevigintillion, 251 duovigintillion, 394 unvigintillion, 333 vigintillion, 65 novemdecillion, 553 octodecillion, 493 septendecillion, 296 sexdecillion, 640 quindecillion, 760 quattuordecillion, 748 tredecillion, 560 duodecillion, 207 undecillion, 343 decillion, 510 nonillion, 400 octillion, 633 septillion, 813 sextillion, 116 quintillion, 524 quadrillion, 750 trillion, 123 billion, 642 million, 650 thousand and 624


my god...

[Anonymous No.4500908]

>>4500892
wouldn't it be 3^250 though?
it can be 0, 1 or both?

[Anonymous No.4500919]

>>4500908
You're a retard. Never respond to a Quantum computer thread ever again.

[Anonymous No.4500922]

>>4500919

[Anonymous No.4500946]

>>4500898
I dunno, 10 years with IBM and many other tech companies having identified this as the future of computing - and hence pouring obnoxious amounts of money into it - could be possible. Also, you can be all major state's military are pushing this too because of it's incredible cryptanalysis potential.
I don't agree with singularitarian bullshit but 10 years seems reasonable with this. Especially considering IBM say they've a working logic gate.

[Anonymous No.4500948]

>>4500908
It can be 0, 1 or anything in between.

[Anonymous No.4500957]

>>4500892
Color me retarded, but my interpretation of this is that a qubit can be in the states [on and on/off] or [off and on/off] due to superpositioning, which again means that 250 normal bits can hold 1^250 = 1 piece of information. What is wrong with this logic?

[Anonymous No.4500964]

>>4500946
Also, I wanted to add that it took only ~30 years to get where we are today with computers which seems ridiculous. Building on all that know-how, industrial base as well as supercomputer simulations makes ~10 years seems like a sensible time-frame.

[Anonymous No.4500969]

>>4500908
There are 2^250 classical configurations of the qubits, in each of which certain qubits are set to 0, and certain qubits are set to 1. The quantum state of the system is an arbitrary combination of all 2^250 configurations. Each configuration gets a complex number whose squared magnitude tells you how likely you are to find the system in that configuration if you measure it. The "2^250 bits of information" blurb is treating all those coefficients as pieces of information. The interpretation is a bit dubious, first of all because those coefficients are complex numbers and not bits, but more importantly, because there's no way to read all that information out, even in principle.

[Anonymous No.4500973]

>>4500948
You're retard, stop posting

It's only 0 or 1. You can only see a cat being dead xor alive and NEVER both. Shit that happen during calculation doesn't matter in the end. You can't store/record a given calculation's intermediate results without getting an answer i.e. you can't find out how dead or alive a cat is until you check its heart.

[Anonymous No.4500985]

>>4500964
I got 30 year of blacksmith experience, therefor I can build me a Quantum computer in 10 years I tell you what!

[Anonymous No.4500994]

>>4500985
I would agree if:
-you had a working quantum logic gate
-you had essentially infinite money
-you had your pic of professors, post-docs, post-grads etc
-you were competing with 1000's of others for what will be a paradigm shift in computing and will make you rich, famous and incredibly good looking.
-AND MANY OTHERS BUT I WON'T WASTE MY TIME BECAUSE YOU ARE CLEARLY A RETARD!!! WAY TO COMPLETELY FAIL AT UNDERSTANDING THE UNDERPINNINGS OF MY POINT IAMNOTEVENMAD.jpg

[Anonymous No.4500998]

>>4500969

Because its the particular permutation that matters. Which bits are set to which state in which order. Its the entire set of 250 that matter, not one at a time

[Anonymous No.4501001]

So I'm still not sure how quantum computers would actually be programmed. Earlier someone said that 250 qbits could store a shitton of conventional data. Could someone post an algorithm or something, to explain how that would actually happen?

[Anonymous No.4501008]

would you be able to overclock those puppies?

[Anonymous No.4501013]

can we use qubits as currency?

[C No.4501029]

ITT: People who don't know shit about quantum computing.

[C No.4501040]

>>4501037
No, I don't know anything about quantum computing

[Anonymous No.4501037]

>>4501029
Care to enlighten us with your vast knowledge?

[Anonymous No.4501057]

Physics and computer science double major... it's the way of the future boys.

[Anonymous No.4501059]

>>4501057
>implying physics has anything to do with quantum theory

[Anonymous No.4501066]

Can I still run my Windows XP with that CPU?

[Anonymous No.4501529]

>>4500904
>MFW while reading that it only just now occurred to me that numbers start counting up using latin prefixes and -illion

>mono = million
>bi = billion
>tri = trillion
>quad = quadrillion
>and so on

God I feel so stupid now.