Libquantum – Simulation of quantum mechanics | Hacker News Reader

Libquantum – Simulation of quantum mechanics

libquantum.de

40 pointsjonbaer11 comments

I find this particularly interesting since I had to implement a quantum simulator using Runge-Kutta numerical integration of the Schrodinger equation in school, for a computational physics class. It was fun to play with (great for making screensavers of waveforms), but I never got so far as learning how to implement quantum computing operations.

There are several programming languages being developed for programming quantum computers (such as QCL[1]), but not a whole lot of awareness of them as far as I know. It'd be really neat to see an open-source QC simulator like this become the basis of a common Quantum Virtual Machine to make quantum programming more accessible to the Interested Hacker.

[1] http://tph.tuwien.ac.at/~oemer/qcl.html

Shameless plug; a while ago I wrote a very similar library [1] using CUDA. Unlike this library, it does not allow for a sparse representation of the state vector; but for maximally entangled states it's pretty efficient (as fast as memory allows). I've tested it up to 28 qubits. But fair warning; it's very much undocumented and a toy.

[1] http://hub.darcs.net/squirrel/quantum

“Can you give me a simple, concrete explanation of how quantum computers work?” http://michaelnielsen.org/blog/quantum-computing-for-everyon...

So... what would actually happen if I used libquantum to run Shor's algorithm?

You would find prime factors.

What else would you expect to happen?

Let me be a little more specific: how would the performance of finding prime factors using Shor's algorithm running on libquantum compare to classical factoring algorithms? I know it would be less efficient, but by how much? How many digits could I reasonably expect to factor in non-geologic time this way? How big a task would it be to code? Is this a reasonable experiment to try to do?

I'm pretty sure it would end up being slower, if only for the overhead. But don't take my word for it, go ahead and run it.

I know it's not too hard to write gates for testing with QCViewer [https://github.com/aparent/QCViewer], (Linux only for the moment, I'm just going to be improving and possibly porting it to OSX)

It would be exponentially less efficient, as the simulation algorithm used represents and operates on all 2^n possible quantum states of n qubits.

If you know nothing about QM, imagine the classical analogue: a classical CPU simulator with exactly one (say) 32-bit register, which, instead of being represented as 32 bits, is represented as 2^32 bits, exactly one of which is set to 1; the only way to determine which bit this is is to scan the entire array.

To expand on the sibling post, you obviously won't find it in faster-than-current factorisation processes because this isn't acutally doing quantum algorithms, but rather emulating them.

What it's useful for is understanding the processes. I found quantum algorithms tough to get my head around at times so this could be a valuable learning tool.

http://qcplayground.withgoogle.com/#/playground/568171774659...

Thanks!!!

Reply on news.ycombinator.com