You are reading without signing in. Sign in or continue as guest to take quizzes and keep XP.
Running on IBM Quantum
Token, instance, queue, minutes: a checklist for a first real-hardware job and how to read the result.
- Set up IBM Cloud credentials safely in QubitLab.
- Estimate queue time and minutes before submitting.
- Judge a hardware histogram against its noisy twin.
- Create an IBM Cloud account and a Quantum service instance (the free Open plan gives 10 minutes of QPU time per month).
- Paste the API key and the instance CRN in Dashboard → IBM Quantum. QubitLab stores the key encrypted and never shows it again.
- Compile first against the backend's fake twin; keep the circuit under 6 qubits and the transpiled depth modest.
- Submit; the Jobs drawer polls the queue. A Bell pair typically takes 2–10 minutes of waiting and a few seconds of QPU time.
The result that matters is not 'did 00 and 11 appear' but how close the histogram is to the ideal one. QubitLab reports the Hellinger fidelity and total variation distance against the exact distribution and against the fake twin; on a Heron machine a Bell pair reaches fidelity above 0.98.
Dry run before the real one
Open the Hardware tab in the Run panel. With no token saved you see the setup steps; with a token you see live backends, their queue depth and calibration. Compile against the fake twin here before you spend minutes.
OPENQASM 3.0; include "stdgates.inc"; qubit[2] q; bit[2] c; h q[0]; cx q[0], q[1]; c = measure q;
Where does QubitLab keep your IBM API key?
Sign in to answer and earn XP; you will come back to this lesson.
A hardware histogram shows 00: 52%, 11: 46%, 01: 1%, 10: 1%. The best summary is…
Sign in to answer and earn XP; you will come back to this lesson.
Save the key once, compile against the fake twin, keep circuits small, and judge results by fidelity against the ideal distribution rather than by whether the right strings appear.
Up next: NISQ reality · Noise and Errors