StandardRB

class StandardRB(physical_qubits, lengths, backend=None, num_samples=3, seed=None, full_sampling=False)[source]

An experiment to characterize the error rate of a gate set on a device.

Overview

Randomized Benchmarking (RB) is an efficient and robust method for estimating the average error rate of a set of quantum gate operations. See Qiskit Textbook for an explanation on the RB method.

A standard RB experiment generates sequences of random Cliffords such that the unitary computed by the sequences is the identity. After running the sequences on a backend, it calculates the probabilities to get back to the ground state, fits an exponentially decaying curve, and estimates the Error Per Clifford (EPC), as described in Refs. [1, 2].

Note

In 0.5.0, the default value of optimization_level in transpile_options changed from 0 to 1 for RB experiments. That may result in shorter RB circuits hence slower decay curves than before.

References

User manual

Randomized Benchmarking

Analysis class reference

RBAnalysis

Experiment options

These options can be set by the set_experiment_options() method.

Options

  • Defined in the class StandardRB:

    • lengths (List[int])

      Default value: None
      A list of RB sequences lengths.

    • num_samples (int)

      Default value: None
      Number of samples to generate for each sequence length.

    • seed (None or int or SeedSequence or BitGenerator or Generator)

      Default value: None
      A seed used to initialize numpy.random.default_rng when generating circuits. The default_rng will be initialized with this seed value every time circuits() is called.

    • full_sampling (bool)

      Default value: None
      If True all Cliffords are independently sampled for all lengths. If False for sample of lengths longer sequences are constructed by appending additional Clifford samples to shorter sequences.

    • clifford_synthesis_method (str)

      Default value: "rb_default"
      The name of the Clifford synthesis plugin to use for building circuits of RB sequences. See Synthesis Methods.

  • Defined in the class BaseExperiment:

    • max_circuits (Optional[int])

      Default value: None
      The maximum number of circuits per job when running an experiment on a backend.

Example

import numpy as np
from qiskit_experiments.library import StandardRB, InterleavedRB
from qiskit_experiments.framework import ParallelExperiment, BatchExperiment
import qiskit.circuit.library as circuits

lengths_2_qubit = np.arange(1, 70, 10)
lengths_1_qubit = np.arange(1, 400, 80)
num_samples = 3
seed = 1010
qubits = (1, 2)

# Run a 1-qubit RB experiment on qubits 1, 2 to determine the error-per-gate of 1-qubit gates
single_exps = BatchExperiment(
    [
        StandardRB((qubit,), lengths_1_qubit, num_samples=num_samples, seed=seed)
        for qubit in qubits
    ]
)
expdata_1q = single_exps.run(backend=backend).block_for_results()

exp_2q = StandardRB(qubits, lengths_2_qubit, num_samples=num_samples, seed=seed)

# Use the EPG data of the 1-qubit runs to ensure correct 2-qubit EPG computation
exp_2q.analysis.set_options(epg_1_qubit=expdata_1q.analysis_results(dataframe=True))

expdata_2q = exp_2q.run(backend=backend).block_for_results()
results_2q = expdata_2q.analysis_results(dataframe=True)

print("Gate error ratio: %s" % expdata_2q.experiment.analysis.options.gate_error_ratio)
display(expdata_2q.figure(0))

print(f"Available results: {set(results_2q.name)}")

Gate error ratio: {'cx': 1.0}
../_images/qiskit_experiments.library.randomized_benchmarking.StandardRB_1_1.png 
Available results: {'EPC_corrected', 'alpha', 'EPC', 'EPG_cx'}

Initialization

Initialize a standard randomized benchmarking experiment.

Parameters:

  • physical_qubits (Sequence[int]) – List of physical qubits for the experiment.

  • lengths (Iterable[int]) – A list of RB sequences lengths.

  • backend (Backend | None) – The backend to run the experiment on.

  • num_samples (int) – Number of samples to generate for each sequence length.

  • seed (int | SeedSequence | BitGenerator | Generator | None) – Optional, seed used to initialize numpy.random.default_rng. when generating circuits. The default_rng will be initialized with this seed value every time circuits() is called.

  • full_sampling (bool | None) – If True all Cliffords are independently sampled for all lengths. If False for sample of lengths longer sequences are constructed by appending additional samples to shorter sequences. The default is False.

Raises:

QiskitError – If any invalid argument is supplied.

Attributes

analysis

Return the analysis instance for the experiment

backend

Return the backend for the experiment

experiment_options

Return the options for the experiment.

experiment_type

Return experiment type.

num_qubits

Return the number of qubits for the experiment.

physical_qubits

Return the device qubits for the experiment.

run_options

Return options values for the experiment run() method.

transpile_options

Return the transpiler options for the run() method.

Methods

circuits()[source]

Return a list of RB circuits.

Returns:

A list of QuantumCircuit.

Return type:

list[QuantumCircuit]

config()

Return the config dataclass for this experiment

Return type:

ExperimentConfig

copy()

Return a copy of the experiment

Return type:

BaseExperiment

classmethod from_config(config)

Initialize an experiment from experiment config

Return type:

BaseExperiment

job_info(backend=None)

Get information about job distribution for the experiment on a specific backend.

Parameters:

backend (Backend) – Optional, the backend for which to get job distribution information. If not specified, the experiment must already have a set backend.

Returns:

A dictionary containing information about job distribution.

  • ”Total number of circuits in the experiment”: Total number of circuits in the experiment.

  • ”Maximum number of circuits per job”: Maximum number of circuits in one job based on backend and experiment settings.

  • ”Total number of jobs”: Number of jobs needed to run this experiment on the currently set backend.

Return type:

dict

Raises:

QiskitError – if backend is not specified.

run(backend=None, sampler=None, analysis='default', timeout=None, backend_run=None, **run_options)

Run an experiment and perform analysis.

Parameters:

  • backend (Backend | None) – Optional, the backend to run on. Will override existing backend settings.

  • sampler (BaseSamplerV2 | None) – Optional, the sampler to run the experiment on. If None then a sampler will be invoked from previously set backend

  • analysis (BaseAnalysis | None) – Optional, a custom analysis instance to use for performing analysis. If None analysis will not be run. If "default" the experiments analysis() instance will be used if it contains one.

  • timeout (float | None) – Time to wait for experiment jobs to finish running before cancelling.

  • backend_run (bool | None) – Use backend run (temp option for testing)

  • run_options – backend runtime options used for circuit execution.

Returns:

The experiment data object.

Raises:

QiskitError – If experiment is run with an incompatible existing ExperimentData container.

Return type:

ExperimentData

set_experiment_options(**fields)

Set the experiment options.

Parameters:

fields – The fields to update the options

Raises:

AttributeError – If the field passed in is not a supported options

set_run_options(**fields)

Set options values for the experiment run() method.

Parameters:

fields – The fields to update the options

See also

The Setting options for your experiment guide for code example.

set_transpile_options(**fields)

Set the transpiler options for run() method.

Parameters:

fields – The fields to update the options

Raises:

QiskitError – If initial_layout is one of the fields.

See also

The Setting options for your experiment guide for code example.