Overview
This guide describes key patterns of behavior and use cases with code
examples to help you migrate code from the legacy backend.run()
interface to use the Qiskit Runtime primitives interface (qiskit-ibm-runtime
package).
- Because
backend.run()
only returned counts, the direct replacement is Qiskit RuntimeSamplerV2
. However, if you used manual processing withbackend.run()
to return expectation values, you can now use Qiskit RuntimeEstimatorV2
instead. - Because both
backend.run()
and the "version 1" primitives are being deprecated, this guide uses only the V2 primitives.
The qiskit-ibm-runtime
package provides cloud access to the IBM QPUs (quantum processing units) through the primitives interface. The backend.run()
interface coexisted with the original (V1) primitives model as the dedicated “direct hardware access” entry point. With the introduction of the V2 primitives interface, the new SamplerV2
class now fulfills that role. Consequentially, backend.run()
is being deprecated, along with qiskit-ibm-provider
, which only exposed the backend.run()
interface.
The Qiskit Runtime primitives implement the reference Sampler V2 and Estimator V2 interfaces found in qiskit.primitives
, and enable capabilities not available with the legacy backend.run()
interface. These capabilities include application of advanced processing techniques for error suppression and mitigation in Estimator, the ability to efficiently sweep between arrays of parameter value sets or observables in both Sampler and Estimator, and access to the new local testing mode. Additionally, Qiskit Runtime lets users run iterative algorithm circuits back to back (session mode) or in collections of circuits without having to re-queue each job (batch mode). This results in more efficient quantum processor use and reduces the time spent running complex computations.
Basic steps to migrate to primitives
Step 1: Determine which primitive to use
When migrating, the key to writing an equivalent algorithm using primitives is to first identify what minimal unit of information your algorithm is based on:
-
If it uses an expectation value of a certain observable with respect to a quantum state (a real number), you will now use Estimator.
An expectation value of an observable could be the target quantity in scenarios where knowing a quantum state is not relevant. This often occurs in optimization problems or chemistry applications. For example, when trying to discover the extremal energy of a system.
-
If it uses a probability distribution from sampling the device, you will now use Sampler.
A probability distribution is often of interest in optimization problems that return a classical bit string, encoding a certain solution to a problem at hand. In these cases, you might be interested in finding a bit string that corresponds to a ket value with the largest probability of being measured from a quantum state, for example.
Step 2: Change imports as necessary
Follow the steps in the appropriate topic to change your import options and other setup information:
Step 3: Replace the call to backend.run
with a call to qiskit_ibm_runtime
.
See these topics for instructions:
- Update code that performs circuit sampling
- Update code that calculates expectation values
- Common use cases (basic, parameterized, and dynamic circuits)
Step 3a: replace any backend.run
options with qiskit_ibm_runtime
options.
See the following topics for instructions:
- Migrate options
- Common use cases Options section
Next steps
- Get started with Estimator
- Get started with Sampler
- Explore sessions
- Run a primitive in a session
- Experiment with the Submit pre-transpiled circuits tutorial