Skip to content

Code to reproduce the benchmarks presented in the qibolab paper.

License

Notifications You must be signed in to change notification settings

qiboteam/qibolab-benchmarks

Repository files navigation

Benchmarks

This repository contains the code required to reproduce the benchmarks presented in Section IV.A. of the qibolab paper.

Note that the main branch of this repository is being updated to newer versions of the Python libraries. For the versions used in the benchmark please check the v0.1.0 tag.

Required libraries

Executing the benchmarks on instruments requires installing the qibolab and qibocal libraries, as well as instrument specific dependencies.

The following versions were used to perform the benchmarks presented in the paper:

Library Version
qibolab 0.1.6
qibocal 0.0.4
qblox-instruments 0.9.0
qm-qua 1.1.6
laboneq 2.25.0
qick 0.2.135
qibosoq 0.0.3

Different versions of these libraries may not support the platforms and runcards provided here.

Setup

The platforms directory provides the platform runcards for the different instruments provided, particularly:

  • zurich for Zurich instruments,
  • qm for Quantum Machines,
  • qblox for Qblox instruments,
  • rfsoc for the RFSoC 4x2 controlled using QICK.

Using these platforms in a self-hosted lab that has access to any of these instruments may require changing some of the setup parameters, such as instrument IP addresses.

The runcards directory provides the qibocal runcards for executing the calibration routines that were benchmarked.

Before executing any routines or programs, qibolab should be configured to read the platform runcards from the correct directory. This can be done by setting the following environment variable:

export QIBOLAB_PLATFORMS=./platforms

If this is run from a different location than this README, the directory should be modified accordingly.

Execution

Benchmarks can be executed using the qq command provided by qibocal, for example:

qq auto runcards/routines_benchmarks.yml --no-update

will execute the benchmarks presented in Fig. 6 of the paper.

Note that running the benchmark on different instruments requires changing the platform key in the qibocal runcards to the corresponding platform name.

Data

Succesful execution of the qq command will create a new folder in the current directory. Note that the name of this folder can be controlled with the -o option in qq.

This folder contains the data acquired for each executed routine as well as a meta.json file with metadata related to the execution. Acquisition, fit and total runtimes for each executed routine can be found in meta.json.

About

Code to reproduce the benchmarks presented in the qibolab paper.

Resources

License

Stars

Watchers

Forks

Packages

No packages published