[scottlocklin]

No, actually they didn't. The Shor algorithm applied to n bits is 72 * n^3 applied gates; for 15 that's 4608 gates; and that's leaving out quantum error correction which would require even more gates. This has never been done.

The result you cite: they don't even consider those early NMR "computers" to be quantum in any meaningful sense. They're avogadro computers.

[q_eng_anon]

ooookay a more recent demonstration: https://arxiv.org/pdf/1804.03719.pdf

you said the number 15 was never factored on a quantum computer - this is false.

You can run as many gates as you want - go ahead and run 4608 gates in qiskit right now - the measurements will be random but you can do it.

IDK where you got that equation for the number of gates but it's probably for the general case - doesn't take into account the fact that different gate sets can be used to reduce the total.

Also confused on the error correction part - the whole point of error correction is to make the coherence time independent of the number of 'gates' in your circuit - so yeah with error correction you get more gates but you also get an effectively infinite coherence time...

[scottlocklin]

>you said the number 15 was never factored on a quantum computer - this is false.

I guess it depends what you mean by "factored" and "quantum computer" -using the generally accepted definitions, the number 15 has never been factored on a quantum computer using the Shor algorithm.

Yes, I am talking about the general case where you don't leave out the gates for 2 and 7 being factors of 15. That's what most people mean by factoring. Stating the answer because you know it already isn't useful. LARPing by running the algo through the "right" gates also isn't useful.

[codesushi42]

You are just plain wrong.

And you failed to consider quantum annealing as an alternative to Shor.

[scottlocklin]

You're right I do fail to consider this, as it's not really an alternative to Shor, because annealing is horse shit that nobody can decide the computational complexity of. Not even D-wave thinks it might be. [1]

[1] https://www.nature.com/articles/s41598-018-36058-z

[codesushi42]

Complexity in what terms? For a classical computer?

Besides, the empirical data shows otherwise. It takes 12 qubits to factor 15. We're up to 53 now.

With quantum annealing, a 20 bit number has been factored with 97 qubits. Not on a real quantum computer yet, of course.

So I have no idea what you are talking about.

[scottlocklin]

> Not on a real quantum computer yet, of course...

Erm, OK. I guess we agree that nobody has factored the number 15 on a quantum computer yet.

Maybe you should read the paper I helpfully linked you above.