We recommend setting up a separate conda environment per vendor; supported are IBM (qiskit), Google (cirq), Rigetti (forest) and Amazon (braket). Each requires the following dependencies.
pip install termcolor matplotlib seaborn networkx
conda install pycairo
The benchmarking program is called runner.py; available commands can be seen with
./runner.py --help
The default output directory is under ./runs, but can be modified.
There are also some bash scripts included that can be used on linux systems to automate some task. The script ./refresh.sh periodically tries to update the benchmarks (e.g. for IBM); it takes as optional parameter a conda environment to launch first.
The default runs can also be invoked via the script
./run-all-benchmarks.sh ibm cloud ibmq_lima
(c) 2021, Johannes Bausch, Arjan Cornelissen, András Gilyén