Slashdot Mirror
IBM Develops Quantum Computer
JSC writes: "IBM has developed a quantum computer consisting of five atoms that work as the processor and memory. It's a nice advance of the state of the art...unfortunately, we won't see them on the shelves for about 20 years." Update: 08/15 06:49 PM by H :Check out the official IBM release - thanks to netMonkey for the update.
----
a beowulf cluster of these? :)
:)
Come on, I deserve karma for posting it on-topic and using my real name
Man, 20 years? well, I bet we will still be arguing whether it should run Linux on Windows by the time these puppy's hit the market...
I am become Troll, destroyer of threads
Can they make it out of Everlasting Gobstoppers instead?
"Linux? Looks like a thousand monkeys at a thousand keyboards to me. Of course, they also throw feces"
This will allow the government to crack 5-bit encryption in fractions of a second! Think of the repercussions for privacy! No longer will the NSA have to brute force their way through the entire keyspace (more than 30 possible keys!)
--
--
Mod up a post Rob doesn't like and you'll never mod again
5 Atoms? We won't see them on the shelve's at all - unless your eyesight's a lot better than mine...
PigPog.
This state would represent both zero and one and everything in between. Instead of solving the problem by adding all the numbers in order, a quantum computer would add all the numbers at the same time.
Does that mean I'll be able to be everywhere at once? Of course that could work both ways -- being everywhere presents a big target...
-------------
The truth is out th- oh, wait, here it is...
There's much more detailed information from IBM Research than was included in the press release. It's interesting that apparently Los Alamos has already developed a 7-qubit computer, it's just that they haven't used it to solve a real math problem yet.
--Chouser
--Chouser
"To stay young requires unceasing cultivation of the ability to unlearn old falsehoods." -LL
I wonder what's going to happen to us programmers. I had heard a while ago in Scientific American that programming Quamtum Computers required a radically different approach than those used today. I wonder how much different it will be, and how us "old-timers" will deal with it.
Let's hope that by 2020, when quantum computing hits the market, Intel will be close to rolling out the PXIII and promising that within 5 years they will make the jump out of x86 architecture. Of course, I'm probably being impractical...
"These are the days that must happen to you." -Walt Whitman
Is it possible that IBM has also developed the world's first Heisenberg compensator?
Don't sneeze!
Just don't sneeze.
134340: I am not a number. I am a free planet!
PigPog's Law: The number of atoms in a quantum computer will double every 18 months.
PigPig Gates says: We'll never need more than 640 protons.
PigPog's Uncertainty Principle: We may know where the computer is or which direction it just blew off the table, but never both at the same time.
PigPog.
Note that the 3dfx video card is 4 of the 5 atoms, and 2 of the 4 are atomic fans. Otherwise, the whole thing would split from the heat. You thought melting was bad -- imagine your computer going critical and wiping out the city! Dang overclockers....
Xentax
You shouldn't verb words.
I would think that a chip like this would be great for PDA's, laptops, just about anything. Really fast, very tiny, and I doubt it consumes any power at all... good deal. I wonder if in even twenty years it will be affordable though. IBM will have to pay off over 20 years of development!
I guess that my big question is this: Are these huge benefits only available in these areas, or could this be used to create a faster general purpose machine?
I know that algorithmic research for these machines is very different than standard CS. Is it just that we've only found good quantum algorithms for these applications? Or is it just that the quantum properties lend themselves to incredible speedups for these specific problems?
I don't know if anyone here can answer these questions, but I'm sure that some of you know more than I do about it! :)
Here's a good starting point for the non-physicist:
Quantum Computation: A Tutorial
And, in fact, has had a working 7qubit computer since March (2000)...
h tm
This article is an easy read with a GREAT summary of the history, applications, and iswsues in quantum computing: http://www.techreview.com/articles/may00/waldrop.
Returned Peace Corps IT Volunteer
For quite some time, it was just a mind game. That was until real algorithms were discovered/invented to take advantage of these curiosities. With powerfully fast algorithms for factoring large integers (the source of encryption's security), searching, etc. they stand poised to change the face of computing. Imagine things such as cracking 1024-bit encyrtption or searching the entire phone-book in one operation.
Of course, the tricky part is to build one. Since they rely on quantum properties, they are easily bumped into a real state. But this is the source of their power too. If one particle can be in two states, then a string of particles can represent every n-bit binary number combination possible!
There are several different ways to go about quantum computing. Some use lasers to cool individual atoms to an energy-level where theyt can be controlled reliably. Others use the bulk-effect of quantum states dtected with nuclear magnetic resonance (ironaically, the caffeine molecule prooves to be particulary useful in this setup).
As of yet, they've been able to do some pretty simple arithmetic with only a few bits of information.
As for how it will change computing and programming, the best guess I've heard is that there might be quantum coprocessors someday (much like the old math-coprocessors). You see, quantum computers are thus far very good at certain kind of operations and not so good at others. This is very similar to traditional CPUs (which suck at factoring numbers in a reasonable amount of time). The two compliment each other.
I knew that information I gleaned while writing that college paper on Quantum Computing would come in handy!
On a similar note, Quantum Encryption is a related field where quantum-entanglement is used to transmit information securely. If someone were to try and eaves-drop on the system the system would collapse into a real state and the information would not be intercepted.
Quantum crypto-cracking (given, say, a 40qubit system) makes cracking traditional crypto insanely easy. IIRC from the defcon quantum crypto talk, it is SQRT(Original keypsace) in difficulty to crack, instead of (Original keyspace).
This means, longer keys, and eventually quantum computing to enable OTPs (One-Time Pads -- yes, with Quantum computing they're possible, and, better yet, functional!)
Returned Peace Corps IT Volunteer
IBM seems to be doing more research and development than anyone else these days yet they don't seem to be focusing one one or two markets. They are already this big bulky monster. Why are they trying to expand like this?
``A quantum computer could eventually be used for practical purposes such as database searches -- for example searching the Web could be sped up a great deal -- but probably not for more mundane tasks such as word processing,''
I really appreciate to see that, though quantum computers could help solving problems, they don't denigrate what came before this.
This will change from "fashion-effects" that make people forget what was sufficient until something new appeared. For example, look at how quick people got rid of the command line when the first GUIs appeared.
--
Trolling using another account since 2005.
I want to calculate solar radiation flux! I want to simulate nuclear detonations!
Why the hell wouldn't you!!?! Video games man!
ID Software are you listening?
Quake7 - The End is Here Real world destruction in real time. You cerbal cord will twitch at the realism and your well ajusted mind will melt in trying to determine what is real, and what is just the video game. (Requires Windows XII, Ultra Linux 7I or MacOS 9.1)
"`Ford, you're turning into a penguin. Stop it.'" -THHGTTG
A Programming Language for Quantum Computers
There is also a good, comprehensive website at
OpenQubit
but it seems to be in need of a new maintainer.
My understanding is that quantum computer simulators allow one to mimic the output of a quantum computer, but without the time speed-up that real quantum hardware would provide. So algorithms can be tested out, slowly, even before powerful quantum hardware is developed. I suspect some problems can also be better expressed in a quantum computing language and would therefore be solved more easily even on classical hardware.
On the subject of simulating quantum physics on classical hardware, in the book The Feynman Processor and in Feynman's own papers it is stated that a classical computer can never perfectly simulate quantum physics. But from the evidence they give it seems merely impractical, not impossible. There can be a huge penalty in the number of steps and time required but no clear reason why a simple quantum physics system could not be perfectly simulated on a powerful classical computer. Anyone have any insight on this problem?
AlpineR
Word: If you're between the ages of 14-18, START STUDYING QUANTUM COMPUTING NOW!!!
Why? Long explanation:
I read half of a book called Introduction to Quantum Computing (can't remember the author, but I bought it at Siggraph'99 -- there was a huge pile of this book in one booth).
Anyway, the book is great. It's almost a step-by-step guide to the math behind quantum computing while still maintaining the physical analogy. I got to the part where they discuss Feynman's method for building a quantum adder (which was merely a trivial demonstration of how to get a QM to do a classical computation).
In chapter 5 or 6, the book starts explaining how to build a Hamiltonion (QM operator function, kinda like a Laplace transfer function H(s)) for the square root of a NOT gate, I realized that anyone who's brain has been fed classical computing concepts based on Turing and Von Neuman is DOOMED to not grok this stuff (or perhaps it's becuase I'm almost 30 and my brain has turned to sand). It's kinda like trying to go from C to LISP.
So kids, that's why I recommend that you start growing the synapses now. Start growing the synapses that will help you understand this stuff before the patterns of classical computing cure in your young gray matter.
(Yeah I love how every reporter goes from: "Fascinating new qubit which is 0 and 1 simultaneously because of spin..." to "...so the qubits add all of the numbers at once to find the asnwer in one step". If you can't explain something in a 5th grade english, you don't understand it.)
---
https://www.accountkiller.com/removal-requested
I want to calculate solar radiation flux! I want to simulate nuclear detonations! I want to solve the traveling salesman problem for 29 billion routes!
You know that's exactly the sort of thing that the old Crays were used for. That was back in the day when "supercomputer" meant something, and these beasts only existed in ones and twos in places like Los Alamos, Sandia, and maybe Exxon.
And back then people thought the exact same things that you are saying now. "Who other than a weapons research lab could possibly use this"? The answer that surprised people is "just about everyone". The Cray-1 may be an inert piece of history now, but it's spirit lives on in our microprocessors. It's not just that modern PC's are as fast as old supercomputers, they are designed like old supercomputers.
Most innovations in computer architecture in the PC/workstation/server area have been taken from the supercomputers that came before. Surely the original researchers never dreamed that all of the complicated methods they were inventing to speed up supercomputers would wind up running some kid's game -- but they have.
Modern systems are blazingly fast, yet people continually feel the need to upgrade. In the PC biz, this seems to be driven by games and MS-bloat. Whatever the case, technology marches on, and people are willing to pay for more power. If you have the transistor budget, why not build a supercomputer on a chip? There's a market for it.
My point (such as it is) is that the hunger for performance shows no sign of stopping. It may seem ridiculous to us that an average person could ever use this much computing power. But bear in mind, that this won't even hit supercomputers for ~20 years. Think what people ~20 years ago would think about the kind of computing power that we use for games today. They would be stunned.
A little historical perspective, that's all...
--Lenny
True, it's easier to break traditional prime-number based ciphers with quantum machines, but there is an effectively unbreakable cipher which can be built off a quantum computer - one that relies off the position of the atoms used as they fly through refractors that "trap" the states, and a system that relies on public-key ideas to keep that atom key a secret.
They cite quantum money as a potential example (an idea developed in the 1970s). There are some truly mindblowing consequences to an unbreakable cipher.
- I don't care if they globalize against free speech. All my best free thoughts are done in my head.
Dropping contact lenses is bad enough. Imagine dropping one of these!!!
"NOBODY MOVE!!!!!"
Bob.
What's the betting that there'll be a story on C|Net tomorrow where some IBM research guy claims that they're considering porting Linux to this thing "just to see how small it can scale down"?
PigPog.
"...unfortunately, we won't see them on the shelves for about 20 years."
You're predicting we'll be able to see 5 atoms on a shelf in 20 years?!
Cool! That's a bigger story than the quantum computer!
This may seem incredibly short-sighted, but what's then to stop us from putting a quantum-computer-on-a-card inside of a traditional computer to handle those sort of problems?
Man, think of the purchasing scams you can pull at your company.
"Bill, I want to talk to you about this 20 million dollar hardware acquisition charge. All we got was this box with foam kernels in it"
"You opened the box?! You fool! You let all the quantum computers out! Now I'll have to reorder them! The board will hear about this!"
Later
Erik Z
Democrats or Republicans. They are both taking us to the same place and they are not afraid of us anymore.
I have to admit first off here that I haven't read up as much as I should about quantum computing. Nontheless, I find myself wondering: what will it be like writing software for a quantum computer? Will quantum machines even be stored-program computers?
Washington, DC: It's like Hollywood for ugly people.
All those cycles...hmmm could it be sentient? Starting to sound a little like Sid 6.7.
>These are the kinds of problems that quantum computers are geared to compute. Not for playing games, not for doing spreadsheets, and definitely not anything for the store shelves.
Whew! I'm glad you figured it all out for us. You saved everyone a lot of time and money by letting us know what what we shouldn't be doing with quantum computing.
Oh, but I think you're wrong on games, I believe you'd be able to program some incredible AI for games like Quake, Or do fantastic universe simulations for games like Elite or privateer.
You're right about the spreadsheets though, who needs spreadsheets when you have have a quantum computer go though all the possible combinations of funding for a company to come up with the best fiscal plan, in about two seconds.
And I'd hate to have one of these things on my desktop, I mean, I'd be playing with it all the time, trying to find out new things that could be done with it. Working together with other cutting edge geeks out there...I mean, what's the point?
Sheesh,
Erik Z
Democrats or Republicans. They are both taking us to the same place and they are not afraid of us anymore.
Maybe this will finally help me solve P=NP...
Right now my solution:
P=NP where N=1 just doesn't seem to work...
~ "When I'm of that age I'm just going to live up a tree."
Burris
Yeah, right. I remember when the Compaq Deskpro 386 came out. "That'll make a great server, but nobody really needs a workstation that fast." So naive. In 20 years, the a.out from hello.c will be 50 megabytes, and you'll need something with a the power to simulate a nuclear explosion just so that MS Word will be able to keep up with your typing speed.
I know. You think I'm joking, or that I'm wrong. But I'm not. Try MS Word 2000 on a Pentium 66 some time and see if it can keep up with your fingers. In 1990, you would have laughed at the idea that computers were going to be slower in 1995 than they were in 1985, but guess what? It happened. Ain't software "progress" great?
20 years from now you're going to be begging on your knees for one of these quantum computers just so you can read your email without falling asleep while waiting for the screen to redraw. And in 2023 your "old" quantum computer will be totally obsolete.
---
As copyright owner of this comment, I authorize everyone to defeat any technological measure which limits access to it.
This part of the book is very "mathematical" and its conclusions must be regarded as true.
Actually, many reseachers, in both physics and AI, do not regard it is true. And the question of "intelligent computers" is still an open one.
Just because something is mathematical doesn't make it true. It may follow from the assumptions, but it is Penrose's assumptions that are in dispute.
See Dennet's Consciousness Explained for an example of a viewpoint opposed to Penrose's.
Steve M
This "quantum computer" is actually a vial full of mole quantities of different quantum computers. You poke at them with radio waves to program them and then use an NMR to read the results, which are a vote by all of the molecules.
That's fine, as long as 10^23 votes is enough to overwhelm any errors. But for a serious number of qubits, the unavoidable chance of quantum bitflip in each atom means that eventually less than one of those 10^23 molecules is in the correct starting state. Perhaps you can solve this by quantum error correction - the algorithms aren't worked out yet - but that multiplies the number of bits needed for a given problem by a factor of (provably) 2 or (probably) 3. Then you have the problem that any operation can only involve closely neighboring bits; to add register A to register C requires huge numbers of operations to shuffle with register B. Finally, to read or write to any given bit, you need a unique frequency to address that bit. With hundreds of bits, only a few can possibly have frequencies that stand out enough. These problems, combined, add one or more factors of N to the resources necessary; it's still polynomial, but...
Quantum dots - single particles, NOT entire atoms, confined electrically to a single quantum state - are more hopeful. Because they can be physically rearranged or put in more complex configurations and still physically addressed on an individual basis, the problems above go away or become more manageable.
Preferential Voting: easy as 1-2-3