search

Qiskit / qiskit

Qiskit is an open-source SDK for working with quantum computers at the level of extended quantum circuits, operators, and primitives.
https://www.ibm.com/quantum/qiskit
Apache License 2.0
5.28k stars 2.37k forks source link

4-qubit Grover on IBM are not good #3410

Closed debfdias closed 5 years ago

debfdias commented 5 years ago

What is the expected enhancement?

I ran a 4-qubit Grover's Algorithm searching for 0000 and 0010 on the IBM backend and did not get even remotely similar results to the simulated runs. I've tried ibmq_vigo, london, essex, none of them are close to the local simulation.

The code for amplification step:

code

And the simulations:

Local:

t_2_local

IBMQ_vigo: t_2_vigo

I've already seen a related issue with 3 qbit Grover, but the results are ok when i run, not that bad as the 4 qbits Grover. Is there any solution for this problem?

1ucian0 commented 5 years ago

Is there any solution for this problem?

If I understand the problem correctly, the short answer is "no, no yet".

The slightly longer answer: I think it might be just the noise in the real quantum chips. You might get slightly better results with optimization level 3. You can also try different backends. However, the current quantum chips are noise and you are trying a relatively deep circuit, so you get "noisy" results. Maybe @kdk, @nonhermitian, @taalexander have more insights.

debfdias commented 5 years ago

Hi! Thank you for the answer. Could you tell me more about this optimization lvl 3?

1ucian0 commented 5 years ago

If you are running execute, you can add the parameter optimization_level=3.

debfdias commented 5 years ago

Thank you very much @1ucian0 . I've tried this but it was not so good for 4 qbit. For 3 qbit i noticed a better performance. Thank you, again.