ISA circuit support for latest Runtime

[woodsp-ibm]

https://docs.quantum.ibm.com/announcements/product-updates/2024-02-14-qiskit-runtime-primitives-update

beginning 1 March 2024, Qiskit Runtime will require that circuits and observables are transformed to use only instructions supported by the system (referred to as instruction set architecture (ISA) circuits and observables) before being submitted to the primitives

I am not sure how much direct impact there is in Finance, if anything, this may be taken care of entirely by changes in its dependents as below.

Finance also makes use of Qsikit Optimization and Qiskit Algorithms for which there is are open issues too on the same subject
qiskit-community/qiskit-optimization#615 and qiskit-community/qiskit-algorithms#164

[L-Choplin]

Hi,

I realize this issue was opened over a year ago, but I’m currently working on option pricing using Qiskit and wanted to ask if there's any update regarding support for V2 primitives, especially SamplerV2().

Right now, most finance-related examples still rely on the older Sampler(). With the Runtime changes and the ongoing migration in other Qiskit libraries, I’m wondering:

• Is there an ongoing or planned effort to integrate SamplerV2() into qiskit-finance?
• Would you recommend sticking to Sampler() for now, or should users begin adapting workflows for V2 primitives?
• Are there any updated examples or branches that already explore this?

Thanks so much for your work on the library — just trying to understand the best path forward for newer projects.

Best,
Louis

[woodsp-ibm]

As of the end of 2023 the Qiskit Applications have new maintainers https://medium.com/qiskit/a-new-chapter-for-qiskit-algorithms-and-applications-5baff541e826 You will note while new maintainers were listed for Nature, Machine Learning and Optimization there was no-one for Finance,

As I noted in the issue Finance depends on and uses Qiskit Optimization and Qiskit Algorithms. Qiskit Algorithms also did not attract a new maintainer so recently Qiskit Machine Learning copied over the parts it depended on and is now independent. There is a PR open in Qiskit Optimization to do the same and also one that was addressing the update to V2 as a PoC (Proof of concept) though it has had nothing done to either of late.

If you want to use the code as-is then you need to stay with the V1 primitives, ie Sampler and not SamplerV2. There is a recent 1.4.2 release of Qiskit that still has these as the applications do not work with Qiskit 2.0 as these have all been removed and it only has the new V2 versions.

If you want to use this code as-is then it needs Sampler. Now you are free to modify the code you are using to have it support SamplerV2 instead if you really need that and it has not been done.

Hope that helps you understand the situation a bit better.

[L-Choplin]

Thank you very much for the detailed explanation, that really clarifies things.

I’ll stick with the V1 primitives for now, and will keep an eye on future updates.