Use Qiskit Runtime SamplerV2 in local mode instead of BackendSamplerV2 in test

[garrison]

This is a draft of something we could do once the next qiskit-ibm-runtime release comes out (we must wait until then because we need Qiskit/qiskit-ibm-runtime#1558). Once that happens, I will likely further iterate on this open another pull request that builds on this to fix some or (ideally) all of the following issues:

• Use qiskit-ibm-runtime in local mode in cutting tutorials #511
• Update cutting tutorials to submit all jobs before calling .result() to obtain them #509
• Use SamplerV2 and EstimatorV2 in documentation #505
• Use Batch mode instead of Session mode in cutting tutorials #535

More info on this migration is at https://docs.quantum.ibm.com/api/migration-guides/local-simulators.

[garrison]

I'm going to add this to the 0.7.0 milestone because it'd be really nice to knock these items off and move to using fake backends and qiskit runtime local mode in the docs as soon as possible. Even if we were to merge this minimal PR before our next release, we could backport the other changes because ckt would already depend on the requisite qiskit-ibm-runtime version. To enable this, though, we'll need a qiskit-ibm-runtime release, either 0.22.1 from the current stable branch or 0.23.0.

[coveralls]

Pull Request Test Coverage Report for Build 8712251243

Details

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage remained the same at 95.438%

---

Totals
Change from base Build 8696380179:	0.0%
Covered Lines:	3452
Relevant Lines:	3617

---

💛 - Coveralls

[caleb-johnson]

Could you please remind me why we need to wait for another release of qiskit-ibm-runtime to knock out all of the issues in the original PR? Is that a dated comment?

EDIT: I see you've opened #552

[garrison]

Could you please remind me why we need to wait for another release of qiskit-ibm-runtime to knock out all of the issues in the original PR? Is that a dated comment?

We need to bump the version of qiskit-ibm-runtime even for this PR to work because it is the first version that includes Qiskit/qiskit-ibm-runtime#1558.

I bumped the version of Aer here, too, so that we can backport #552 (if it doesn't land in time for the release) without needing to further bump any versions on the stable branch. Aer 0.14 is required there because that is when EstimatorV2 was introduced.

[caleb-johnson]

This looks great, thanks!

Line 51 in the test uses GenericBackendV2. Should we make a test explicitly testing against that backend and use AerSimulator everywhere by default? The only places we would need to switch would be here and Tutorial 2, cutting for depth.

[garrison]

Line 51 in the test uses GenericBackendV2. Should we make a test explicitly testing against that backend and use AerSimulator everywhere by default? The only places we would need to switch would be here and Tutorial 2, cutting for depth.

Actually, in #552 I am transitioning to using Qiskit Runtime in local mode with fake backends in all the tutorials, so I think that, the test you mention, plus AerSimulator in this PR's test covers all our bases.

We can always continue this conversation at #552, too.