pyqbench 0.1.2

https://arxiv.org/abs/2304.00045

PyQBench is a package for benchmarking gate-based quantum computers by estimating how well they can discriminate between two von Neumann measurements. PyQBench is built around the Qiskit ecosystem and its configuration is driven by YAML files describing the experiment scenarios and backends to be used.

Installation

PyQBench can be installed from PyPI using pip:

pip install pyqbench

Quickstart

The most basic way to use PyQBench is by using its CLI. For more advanced usages see PyQBench's docs.

PyQBench's CLI can only run experiments using the parametrized Fourier family of measurements. Here's a basic example of how it works:

Step 1: preparing configuration files

The first YAML configuration file describes the experiment scenario to be executed:

type: discrimination-fourier
qubits:
    - target: 1
      ancilla: 2
angles:
    start: 0
    stop: 2 * pi
    num_steps: 3
gateset: ibmq
method: direct_sum
num_shots: 100

The second file describes the backend. The precise format of this file depends on the type of the backend, here's an example for Qiskit's IBMQ backend:

name: ibmq_quito
asynchronous: true
provider:
    hub: ibm-q
    group: open
    project: main

IBMQ backends typically require an access token to IBM Quantum Experience. Since it would be unsafe to store it in plain text, the token has to be configured separately in IBMQ_TOKEN environmental variable.

Step 2: running the experiment

After preparing YAML files defining experiment and backend, running the benchmark can be launched by using the following command line invocation:

qbench disc-fourier benchmark experiment_file.yml backend_file.yml --output async_results.yml

The benchmark command will submit several batch jobs that will run asynchronously. It is non blocking, meaning that it won't wait for the jobs to finish.

You can check the status of the submitted jobs by running:

qbench disc-fourier status async_results.yml

Step 3: resolving asynchronous jobs

To get the actual results from the submitted asynchronous jobs, you can use resolve command.

qbench disc-fourier resolve async-results.yml resolved.yml

This operation is blocking and will wait for all the results to be available.

Step 4: tabulate results

Finally, the obtained results can be summarized into a table.

qbench disc-fourier tabulate results.yml results.csv

Here's what the result looks like:

What else can PyQBench do?

The above quickstart guide does not cover:

• Running the experiments in synchronous (blocking mode)
• Using error mitigation via M3Mitigation suite
• Using user-defined measurements instead of the default Fourier

Refer to PyQBench's documentation for further reading.

Authors and Citation

PyQBench is the work of Konrad Jałowiecki, Paulina Lewandowska and Łukasz Pawela. Support email for questions dexter2206@gmail.com. If you use PyQBench, please cite as per the included BibTeX file.