A conventional, binary processor can perform some incredibly complex calculations at blistering speeds, measured in FLOPS: floating point operations per second.
The fastest "conventional" processor array running today is China's Tianhe-2, which has achieved speeds of 33.86 petaFLOPS. This is 33 x 1015 floating point operations per second. There are plans to achieve exaFLOP speeds within the next five years or so; 1 x 1018 floating point operations per second.
Are you still with me? Okay, take a breath.
These projects are usually undertaken by governments, and can be quite expensive. James Bamford reported that the U.S. National Security Agency, for example, is said to have requested an exaFLOP capable computer by 2018.
We can only wonder what they'll use it for...
But quantum computing takes all this output, all of this incredible performance, and turns it on its ear.
Harnessing the Building Blocks of the Universe
Zero or one, on or off, go or stop, yes or no: until now, this has been the traditional approach to computing - strictly binary.
But we sit on the cusp of a complete rethink of computing, the dawn of quantum computing.
In this arena, it's a question of zero and one, on and off, go and stop, yes and no - all of these, or none, all at once. Scientists will have the ability to target and control the spin of quantum particles, giving rise to unprecedented computing power.
If that seems difficult to get your mind around that possibility, there's a good reason for it. Quantum computing relies on some of the most fundamental - but complex - properties of the universe, particles at their smallest level.
Quantum particles live in a whole range of states, quantum superposition.
If a system might be in any one of a number of different configurations, then the most "general" state is some combination of all of these different possibilities.
Only some of the information about these states is observable to us. We can know the speed of a particle, for instance, but not its position, or its position, but not its speed, and so on. This is a vast oversimplification of a phenomenon called the Heisenberg Uncertainty Principle.
It might be easier to think of it as a situation where anything that could happen does happen. And when you consider that these different outcomes could be recorded with at least some certainty, you get a hint of the potential of quantum computing.
In fact, achieving this sense of certainty is one of the biggest hurdles facing quantum computing... but for how much longer?
Not much longer at all, it turns out. According to the Christian Science Monitor, scientists in Australia and the United States have solved a key problem: how to apply "spin" to each qubit effectively, and eliminate the effect of other spinning particles on the processing of data.
Scientists believe they've worked out most of the kinks relating to how reliable the results are, and companies like Lockheed Martin Corporation (NYSE:LMT) are cautiously optimistic that we're on the verge of a new era:
"This is a revolution not unlike the early days of computing... It is a transformation in the way computers are thought about."
Ray Johnson, chief technology officer of Lockheed, characterized quantum computing this way in an interview with The Hindu.
Everything, All At Once
If brought to maturity, quantum computing will be able to solve in minutes what took weeks or months before. We may live to see the end of conventional medical testing.
If a computer can analyze all possible combinations and report on all possible outcomes, why test a drug at all? How many billions of dollars could be saved if we knew, more or less instantly, how new technology was going to perform as soon as it could be modeled?
This technology, the entire field, is not without its problems, not without its unanswered questions. There have been false starts before, and there are still legitimate questions as to the validity of all the data being produced.
Canadian company D-Wave is a private quantum computing research company based in Burnaby, British Columbia. The company has attracted the attention of a diverse array of financial and government heavy hitters. Jeff Bezos of Amazon, Goldman Sachs, and the Central Intelligence Agency are all lined up to hear what D-Wave has to say.
They claim to have the world's first quantum computer, D-Wave One, ready for sale. Its processor, code-named "Rainier," is said to be the world's first 128-qubit (one qubit is the basis of quantum computing, much like the bit for conventional processors) processor. Don't plan on setting a D-Wave One up on your desk just yet. The cost is a hefty $10 million per unit.
Then again, you may not want old "Rainier" anymore, now that "Vesuvius" is on offer. The "second generation" processor packs a formidable 512-qubit processor.
But bigger fish are interested in what D-Wave has to offer. Lockheed Martin, Google Inc (Nasdaq:GOOG), and NASA, have been working with the company on some very interesting quantum computing applications, in some cases they've combined quantum computing with a cloud of traditional computers.
Google, for instance, has been working with D-Wave to create recognition algorithms for streets and landmarks. This is an important problem in creating self-driving cars.
NASA and Lockheed are interested in the prototyping and testing benefits that a quantum computer could provide. Google and NASA's Universities Space Research Association have founded a lab devoted to quantum computing research.
As incredible as processors like "Rainier" and "Vesuvius" are, they represent the first baby steps in this field.
How Quantum Computers Work
- Christian Science Monitor:
Scientists Clear Quantum Computing Hurdle
- The New York Times:
Google Buys A Quantum Computer
- The Hindu:
Quantum Computers Turn Mainstream