Quantum computing firm D-Wave bags $17m more in funding D-Wave Systems has attracted a lot of criticism from computer scientists over claims it has developed a way to create a marketable quantum computer. But whether its technology will ever be viable outside of a laboratory setting or not, investors seem to be eating it up. The company announced this week that it received $17m in …
I think you've gotten the idea behind quantum computing a bit wrong. Qubits aren't just able to store 00 01 10 11 instead of only 0 or 1. If that were the case then they'd be nothing special.
The reason that Qubits are so interesting is because you can use them to solve problems that are impossible for "normal" computers to achieve in a practical timeframe.
For example, if you want to figure out which two prime numbers multiply together to make 65, a normal computer has to do this by trial and error, for example see if 65 can be divided by 2 (no), then see if it can be divided by 3 (no), divide by 5 (yes) -> ok then the answer is 5 * 13. When the numbers are very big, this takes billions of years even with very powerful computers.
Now the reason a quantum computer could be useful for this problem is because it can effectively try all possible combinations of prime numbers at the same time and immediately find the single correct answer. The Qubits are the reason it can do this.
There are a lot of these trial-and-error problems in practical situations that could be revolutionised by Quantum computing. Cryptography is the most obvious one but there are plenty of others.
Jesus, that attempt at explaining a quantum computer had functional equivalence with the operation of an aspiration device (in fairness, I don't understand them either, I need free time to get through my Schaum on Quantum Mechanics!! Help!) But anyway, here's another link into D-Wave controversy which has a must-read lively discussion in the comment section:
http://www.scottaaronson.com/blog/?p=291
Note that quantum computers can be faster than classical ones - _if_ you can find a way express your problem using quantum gates in the first place. Are there any practical problems that fit that criterion except factorization and fast table lookup? Maybe protein folding? They don't seem to be able to help you crack NP-hard problems, so no cigar.
I'm no quantum computing expert, but I thought one qubit could be used to simultaneously solve two solutions in one moment by being in two states at the same time. This article says one qubit represents 4 states, which is correct?
From the article:
1 qubit = 4 states
2 qubits = 8 states
4 qubits = 16 states
6 qubits = 24 states
This is clearly wrong as the author is just multiplying the qubits by four. I'm not going to lecture readers as to how to properly compute exponential growth, suffice it to say this is not it. How did that slip?
A single qubit can be either 0 1 or a superposition of both values.
So one qubit can actually only occupy 3 states and not four.
The third state does not exist in a classical sense and therefore is not a valid output from a quantum system. A measure of the quantum state of the qubit produces a 1 or 0 not the superimposed 1 and 0.
A system with n qubits can perform 2^n calculations at once.
For a 2 qubit system this relates to 4 possible values at once.
For a 3 qubit system this relates to 8 possible values at once.
and so on.
Well, actually "1 qubit" is a shorthand for a 3-vector of length 1 (one endpoint lying on the origin, and one on the sphere with radius 1), whereby you encode:
1) Phase (an angle between 0 and 2Pi, the vector rotates around the z axis)
2) Superposition of 1-ness and 0-ness, which determines whether the qubit gives a 1-bit or 0-bit when measured. Full 1-ness if the vector points in the positive direction of z, full 0-ness if the vector points in the negative direction of z.
Your computer becomes interesting if you entangle the qubits so that after some time, a measurement of a qubit array will yield interesting solutions with high probability.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
