Published May 18, 2020 | Version 1.0

Application-Motivated, Holistic Benchmarking of a Full Quantum Computing Stack: Experimental Data

  • 1. University of Edinburgh
  • 2. Cambridge Quantum Computing Ltd
  • 3. IBM T.J. Watson Research Center
  • 4. Cambridge Quantum Computing Ltd, University of Strathclyde

Description

Full experimental dataset for the publication "Application-Motivated, Holistic Benchmarking of a Full Quantum Computing Stack". The archive `application_motivated_benchmarks.zip` contains the following files and directories:

- uncompiled_log.csv

Gives IDs for the uncompiled circuits initially generated for use in our
experiments, along with the properties of the circuits.

- properties_log.csv

Gives IDs for device property files, along with the device and the time at which
they were collected.

- compiled_log.csv

Gives the calculated figures of merits for the compiled and run circuits.
Compiled circuits are identified by the ID of the uncompiled circuit, the
compilation strategy used, and the device compiled onto. Device property IDs at
the time of compilation and run are given.

- circuits/

Contains a subdirectory for each uncompiled circuit. Each subdirectory has files
of 2 forms.
    
    - uncompiled.qasm is the uncompiled circuit.
    - files of the form 'strategy'_'device'.qasm are the compiled circuits.

- data/

Contains a subdirectory for each uncompiled circuit. Each subdirectory has files
of 3 forms.

    - prob_vector.csv contains the ideal output probability distribution.
    - files of the form 'strategy'_'device'.csv contain the shot counts for
      each compiled circuit when run on the real device.
    - files of the form 'strategy'_'device'_simulated.csv contain the shot
      counts for each compiled circuit when run using a classical simulator
      with noise model build from device properties at the time of the real
      run.

- device_properties/

Contains json files detailing device properties for each device property ID.

 

Files

application_motivated_benchmarks.zip

Files (72.0 MB)

Name Size Download all
md5:8322941db0d22d352846b89262251be3
72.0 MB Preview Download