-
Notifications
You must be signed in to change notification settings - Fork 2.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add a BackendV2 version of FakeOpenPulse2Q #8712
Comments
It seems useful to me, but I don't have the full details. In that vein, having a V2 version of I would be against adding methods to any prospective |
No,
|
Let me leave this for @mtreinish when he's back next week - I'll probably just get more things mixed up and not help anything. |
Here the real IBM V2 backend reports To me what @wshanks is proposing for IBM fake provider seems reasonable upgrade. Perhaps, it would make more sense to have What I'm not really clear about V1 and V2 model is the future migration plan of |
Right, From the qiskit-terra and general qiskit user perspective a |
This is what I misunderstood and this approach is sensible. One the other hand, in Qiskit Experiments, we need to write unittests with these fake provider's backends, and we must expect the same code works for the real backends from the IBM provider. So the change we will make locally should immediately propagate to IBM backends, and the updated model should be still architecture agnostic (note that we are talking about the low-level hardware configuration). The important point, which is currently missing in the V2, is the field to describe the hardware configuration (e.g. channel frequency, source power, amplifier gain, etc... -- and available configuration depends on provider/access level of the client), and this should be a separate field from calibrations since such a hardware configuration is not gate dependent. I think For example, drive_freqs = backend.target.metadata["drive_freqs"]
meas_freqs = backend.target.metadata["meas_freqs"] Any thoughts? @wshanks |
What should we add?
Would it be reasonable to add a
BackendV2
version ofFakeOpenPulse2Q
? What would need to be added besides the class? There are not really tests for the V1 version, but it is used in several tests.I made a hacked V2 version in qiskit-community/qiskit-experiments#900 for reference. Another option might be to wait for #8611. I am not sure about timing and priorities.
Also, a question that came up in qiskit-community/qiskit-experiments#900 and Qiskit/qiskit-ibm-provider#387 -- could the V2 version of
FakeOpenPulse2Q
have adefaults()
method? Could the other fake V2 backends have it as well (seefake_pulse_backends.py
in qiskit-community/qiskit-experiments#900).I could work on this but I wanted to see if it would be acceptable or if this is not moving in the same direction as
BackendV2
and should just be hacked around outside of terra (like in qiskit-community/qiskit-experiments#900) until it is not needed.The text was updated successfully, but these errors were encountered: