Over the last month or so, Quantum Computing became the hottest topic around when Google boldly stated that its D-Wave 2X Computer was a 100 million times faster than today's computers.
Sure this all sounds fantastic, if a little fanciful, so surely there has to be a catch right? Well, it's more than that. The D-Wave 2X is a 512 Qubit PC which is kept at a fraction of Absolute Zero temperature to achieve maximum performance. On top of that it requires at least a 10 square meter room. The D-Wave One was priced at US$10,000,000. Imagine the price of D-Wave 2X? Bill Gates would probably have to think twice before whipping out his chequebook.
So just what are the basics of Quantum Computing and where exactly does Quantum Computing have an edge against our traditional PCs?
- What is Annealing?
Annealing is a technique of finding the global minimum and maximum of a given function, through an algorithm. Currently there are two types of Annealing which we'll be discussing: Simulated Annealing, which is used in everyday PCs, and Quantum Annealing, which tech experts have touted as the future of computing.
- Simulated Annealing:
Simulated Annealing is an algorithm used to solve totally unconstrained and bound-constrained problems. The method follows the model used in metallurgy and crystallization whereby thermal jumps are used to find global maximum to any given problem while using the minimum energy possible. In metallurgy this is achievable by lowering the temperature slowly. Basically you search for a good solution to a problem rather than the best one and gradually make it the best one, probing for answer.
Think of it like a map of a giant land scape, filled with mountains and valleys. On it you are trying to find he lowest possible point. With simulated annealing, you are scanning over that entire landscape bit by bit, before you return the complete data set and have your answer.
- Quantum Annealing:
Quantum Annealing is an algorithm which implies the use of Quantum mechanics in an adiabatic system. An adiabatic system is one in which there is no external influence of pressure, temperature and so forth. Unlike Simulated Annealing, which has to go through several jumps to find the global minima or maxima, Quantum Annealing basically cuts straight through these hurdles to find the minima. Observe the following picture from Wikipedia:
So using the landscape analogy above, quantum computing can simply cut through the mountains and valleys, arriving straight at the correct answer in a far quicker process.
A Qubit is a shortened term for a Quantum bit. Quantum Bits, like ordinary Bits, can be 1 or 0. At first look glance it seems totally similar to a regular Bit, but a quantum bit is also capable of being both 1 and 0 states at the same time. It essentially means a qubit can exist in more than one state at any given time, and the more Qubits there are the more simultaneous states it can be in. This, along with quantum tunnelling, means quantum computers can work on manipulate all sorts of bits and combinations of bits simultaneously.
- Where Does Quantum Computing Have An Edge Against Traditional Computing?
First of all observe the above diagram. You will clearly notice that Quantum Tunneling is using a shorter path to solve a given problem while Simulated Annealing has to do periodic thermal Jumps to avoid the traps. Quantum Tunneling simply penetrates though energy barriers without any additional cost whatsoever.
Coming to the D-Wave 2X, Google and NASA tested this computer and a traditional one with 3 or 4 problems, such as one involving D-Wave against Simulated Annealing, Quantum Monte Carlo etc, or order to find out which was the quickest.
In the first part, D-Wave 2X outclassed a traditional computer by 1.8 x 10^8 (180,000,000) times in a problem which was designed for the specific failure of Simulated Annealing. In this part, Single Core Performances of both computers were compared. However, due to the problem being exceptionally complex, the Traditional PC would reset several times and distribute the load across various cores to find a solution to the problem. Quantum Computing had no such issue despite the fact that not all Qubits in D-Wave were properly calibrated.
Similarly when D-Wave was run against Quantum Monte Carlo, it had an advantage of 10^8 at max load. Further comparison with other Classical Solvers and as well as an unverified claim at this time of writing shows that Quantum Computing clearly possesses a substantial advantage against Simulated Annealing when it comes to specific types of problem solving computing.
- Promising Factors and Impacts of Quantum Computing
With much power comes great responsibility!
The D-Wave 2X regardless of its size, maintenance and other factors offers extraordinary power compared to any other computers to date. Utilization of this power for something other than solving complex binary problems could lead to huge advance in scientific research. Complex problems related to science, medical or other fields could be solved in a matter of minutes compared to hours, days or even years with other common annealers.
For general purpose computing and equation work the D-Wave 2X isn't even flexing its muscles. It can only be fully utilised in the creation of artificial intelligence machines and other complex problem solvers.
If the potential of Quantum Computing could be unlocked for more general usage, game design, in-game artificial intelligence, or virtual reality could establish it at an entirely new level. As Quantum Computing can be used to find the minimum solution to a problem, it can even be used towards writing bug free code or optimising performance. It's far easier said than done though, particularly with a $10 million price tag attached. It's unlikely we'll even begin seeing quantum computers tested in these situations for the next decade or so.
As well as the brute force performance, more conservation of energy will be possible again due to the process of quantum annealing. The computers developed by Google and NASA use an adiabatic system meaning the surrounding temperature has zero effect on the working of computer which is working in a fraction of Absolute zero.
- Controversies Of D-Wave 2X
- Physicists are not totally in agreement regarding the Quantum Computing techniques being used in Google's Quantum Computer. Some argue it isn't true quantum computing, with a number of non-quantum effects being used.
- A few believe that the results obtained from these experiments are not as exceptional as Google claims. Perhaps they want Quantum Computing to outclass Simulated annealers and in the process introduced a new prefix to measure this advantage.
- Claims of D-Wave being 100 million times faster could be exaggerated, which some claim was a trick from D-Wave's marketing department.
And as you would predict, Google and family have thrown out more research to back up the claims of Quantum Computing in D-Wave. Still, some remain unconvinced and demand the Simulated Annealers should be given more friendly algorithms.
Right now it's fair to say that Quantum Computing is indeed a huge leap from our common conventional computers such as the Intel Xeon E5-1650 CPU @ 3.20GHz which was used to determine one of the tests. Google hasn't specifically stated the performance boost when comparing multi-cores as well as the optimization of programs on Quantum Computers either.
Quantum Computing may also have an effect on the programming languages we use nowadays, with common thinking being it will make them more efficient.
Right now, this is clearly the first generation of Quantum Computing, so it's still very much in its infancy. By the time the 4th or 5th generations of Quantum Computing roll around, we could well have these computers in smaller or less sophisticated forms in our homes.
Well that's a lot of read GD'ers. I hope you all got a little grasp on what Quantum Computing is, and just how complex a topic it is to broach. Leave us your thoughts in the comments as well as check out the references enclosed below for further information.
- Quantum Annealing Research Paper Related To Finite Range Tunneling by Vasil S. Denchev, Sergio Boixo and others.
- How "Quantum" is the D-Wave Machine? by Seung Woo Shin, Graeme Smith and others.
- D-Wave's Year of Computing Dangerously by IEEE Spectrum
- Why Is There So Much Controversy Around Quantum Computing? - Quora
- Defining and Detecting Quantum Speedup by Troels F. Rønnow, Daniel A. Lidar and others.