Tony,
You folks selling these yet?? I want one!! But, Oh goodie, now I guess I will have to back to school in a few years. <gg>
Have a good evening!!
Take Care
Ann
usatoday.com
07/13/99- Updated 09:50 PM ET
Beyond the PC: Atomic QC
Quantum computers could be a billion times faster than Pentium III
By Kevin Maney, USA TODAY
Around 2030 or so, the computer on your desk might be filled with liquid instead of transistors and chips. It would be a quantum computer. It wouldn't operate on anything so mundane as physical laws. It would employ quantum mechanics, which quickly gets into things such as teleportation and alternate universes and is, by all accounts, the weirdest stuff known to man.
This quantum computer would be a data rocket. It probably would do calculations a good 1 billion times faster than a Pentium III PC. It would be able to search the entire Internet - imagine how much will be on the Net in 2030 - in a blink, and could break any cryptographic security code ever invented, no doubt making the CIA very, very nervous.
Sound like science fiction? It's not. Over the past year, quantum computers have become a serious contender for What Comes Next - after Moore's Law takes the current architecture of transistors mounted on microprocessors as far as it can go in increasing processing power.
Quantum computers do calculations using atoms instead of computer chips. The first quantum computers are still rough, expensive, one-shot science experiments. But since last year they have been built and have shown that the science works. Labs at places like the Massachusetts Institute of Technology and Oxford University are pumping up quantum computer projects. Companies such as IBM and Hewlett-Packard are leaping in. The federal government, which is both worried about and intrigued by quantum computing, has set up one of the most well funded quantum computing labs at Los Alamos National Laboratory.
“This area has gone off like a big bang. It's breathtaking,” says Stan Williams, head of Hewlett-Packard's labs. “The potential is so huge and it would be so disruptive, it could completely change the way at least some computing is done.”
The caveat, though, is that quantum computing is tremendously hard in practice and in theory. In practice, creating a situation where atoms do calculations and give results challenges even the best scientists.
On the theory side, quantum mechanics delves deep into areas that are nearly unthinkable. For instance, it's possible that a quantum computer holds an infinite number of right answers for an infinite number of parallel universes. It just happens to give you the right answer for the universe you happen to be in at the time. “It takes a great deal of courage to accept these things,” says Charles Bennett of IBM, one of the best-known quantum computing scientists. “If you do, you have to believe in a lot of other strange things.”
The result is that practical quantum computing is still decades away. Scientific efforts today are the quantum computing equivalent of Ben Franklin flying his kite in a lightning storm to test theories of electricity. The next step for the labs is to figure out how to control and bottle this incredible force.
Quantum basics
For a non-scientist, understanding how quantum computing works is darn near impossible. But the basics are worth taking a stab at.
Quantum computing's starting point came when physicists realized that atoms are naturally tiny little calculators. “Nature knows how to compute,” says MIT's Neil Gershenfeld, who with IBM's Isaac Chuang built the most successful quantum computer yet. “We just didn't know how to ask the right questions.”
Atoms have a natural spin or orientation, in the way a needle on a compass has an orientation. The spin can either be up or down. That coincides nicely with digital technology, which represents everything by a series of 1s and 0s. With an atom, a spin pointing up can be 1; down can be 0. Flipping the spin up or down could be like flipping the switch on and off (or between 1 and 0) on a tiny transistor.
So far so good. But here's one of the weird parts, which also is the source of quantum computing's great power. An atom, which is not visible to the naked eye, can be both up and down at the same time until you measure it. The act of measuring it - whether using instruments or a powerful microscope - forces it to choose between up or down.
Don't ask why it works that way. It's part of quantum mechanics, which is a set of laws - like Einstein's theory of relativity - that govern the universe. In this case, the laws govern the tiniest objects in the universe. Quantum mechanics is entirely unlike anything in the world of ordinary experiences. The laws are so bizarre they seem made up. Yet they've been proven time and again.
Since an atom can be up and down at once - called putting it into a superposition - it's not just equal to one bit, as in a traditional computer. It's something else. Scientists call it a qubit. If you put a bunch of qubits together, they don't do calculations linearly like today's computers. They are, in a sense, doing all possible calculations at the same time - in a way, straddling all the possible answers. The act of measuring the qubits stops the calculating process and forces them to settle on an answer.
Forty qubits could have the power of one of today's supercomputers. A supercomputer trying to find one phone number in a database of all the world's phone books would take a month, Chuang says. A quantum computer could do it in 27 minutes.
Different kind of spin
Another aspect of quantum mechanics could prove important to computing. It's called entanglement. An outside force acting on two atoms can cause them to become entangled. Wherever they are in the universe - even light-years apart - they will stay entangled. Their spins are in all positions at once. But the instant one entangled particle is observed its spin goes one way. At that same instant, the spin of the other particle, or atom, locks in the opposite way.
In a sense, it's communication. If you see an entangled particle lock in an up position on this end, you know it is down on the other end. And since it happens instantaneously, it seems to defy laws about the speed of light. In fact, theories about entanglement have led scientists to believe there are ways, however improbable, to do Star Trek-style teleportation.
More realistically, entanglement could be a way to speed up computing. Even today's computers are nearing a point at which their speed is being limited by how fast an electron can move through a wire - the speed of light. Whether in a quantum or traditional computer, entanglement could blow past that limit.
The concept of programming the software for a quantum computer isn't any less strange.
To program a quantum computer, you wouldn't use the step-by-step logic of today's computers. You'd want logic that used the peculiar properties of qubits. That's what Lov Grover of AT&T's Bell Labs did when he invented an algorithm, or mathematical program, that uses quantum computing to search databases. He describes it as similar to dropping multiple pebbles in a pond so that the waves cross and interact in a particular way.
Grover's algorithm sets up multiple paths of computations, so that waves of results - all happening at the same time - start interfering with each other. “You get the undesired answers to cancel out,” Grover says. “The right answers interfere constructively and add up.” It's kind of backward computing: You assume the computer already knows all possible answers and it has to find the right one.
Theory to practice
For most of the 1990s, people like Grover, Bennett and Gershenfeld have operated in a world of only theory. The tough question, though, has been: How do you make a quantum computer work? Lately, there have been some answers.
The Gershenfeld-Chuang quantum computer looks nothing like a computer. It looks more like a nuclear toaster. The most intractable problem of quantum computing is that the inner workings of the device, the actual calculating atoms, have to be completely isolated from their surroundings. Any interaction with even a single other atom or particle of light makes the particles choose a spin direction, polluting the results.
And yet, if you're going to program a quantum computer, put in data and get out a result, you have to interact with the atoms somehow.
A year ago, Gershenfeld and Chuang, along with scientists at Oxford and the University of California at Berkeley decided to use a nuclear magnetic resonance machine, or NMR, often used for brain scans. They filled a test tube with chloroform fluid, which is made of molecules composed of carbon and hydrogen atoms. The scientists lowered a test tube filled with chloroform within a magnetic coil that emits controlled magnetic pulses.
The atoms in the chloroform do a dance with their spins according to a pattern dictated by nature. The NMR pulses nudge some atoms in the dance, indirectly affecting the spins of other atoms. That way, the carbon spins can be programmed without any contact, allowing them to perform as quantum computers.
The dancing produces a slight warbling in the magnetic field. By measuring that warble, scientists can read the result of the quantum calculations. “NMR kicks the spins at precise intervals,” Chuang says. “The sequence of flips is the program.”
The program contained Grover's algorithm. It did a simple search, finding one item out of four in a single step. (A regular computer would take two or three tries.) Gershenfeld and Chuang ended up creating the first 2-qubit quantum computer. It cost about $1 million. They have since created a 3-qubit computer.
There are other ways to make quantum computers. A group in Australia is trying to make one that doesn't use liquid. Another group has tried something called ion traps, which essentially makes one quantum computing particle at a time.
The labs around the world are pushing the science hard. “Everyone wants to be the first group to add the next qubit to the system,” says H-P's Williams. “The group that comes up with four - that'll be a big deal.”
Pushing physical limits
For a long time, quantum computing will be a big deal only to scientists. Yet computer companies have more than a passing interest. As Gershenfeld explains, if the transistors on computer chips keep getting smaller and more packed together at the same pace as over the past 25 years, by about 2020 the width of a wire on a computer chip will be the width of a single atom. “Every physical parameter you can think of hits bottom,” he says. “It may even come sooner.”
At that point, computers using today's chip designs would stop getting rapidly better, faster and cheaper. Something will have to take up the slack. Quantum computing looks like an attractive option - because of its potential power, because it seems to work and because the supply of raw material is more endless than silicon. “It's the biggest untapped resource in the universe,” Gershenfeld says.
Another group watching closely is the cryptography crowd. “It has the potential of making a lot of our stuff obsolete,” says Bruce Schneier of computer security experts Counterpane Systems.
Cryptography, he says, is based on extremely hard factoring problems, which are nearly impossible to do on a traditional computer in any reasonable amount of time. “Quantum computing allows you to use an infinite number of hammers at the same time” to whack apart a cryptographic code, Schneier says. “Then it's no longer a hard problem.”
In that sense, quantum computing could be a can of worms. Credit card transactions, stock trades, military messages, government communication - all depend on cryptography to stay secure. The CIA, the military and the Department of Energy have been funding research and keeping tabs on other work in the field. They don't want to be caught flat. “Any technology like this has the potential for great benefit and great harm,” Williams says.
Will it become an everyday product? Maybe someday. H-P envisions a quantum computing attachment. Most of your PC would be of the traditional variety, but you might plug in a quantum device when it's time to do the things quantum computing does best, such as search or cryptography. Other experts, including Gershenfeld, see a day when you could walk into Wal-Mart and pick up a liquid-filled desktop quantum machine.
The only outlook that's clear is that quantum computing keeps getting more plausible and electrifying. “Things we thought were limits are not,” IBM's Bennett says. “In some ways, we've been overly pessimistic.”
|
