Quantum Benchmark Suite for IBM’s Backend based on SupermarQ

Abstract—The benchmark of Noisy Intermediate-Scale Quantum
(NISQ) computers plays a major role in the research and
development of quantum devices and algorithms. Our study
aims to investigate a single metric of benchmarking for the
Noisy Intermediate-Scale Quantum device. We use state-of-theart
applications to generate a device-score according to our
proposed formula. Our study utilizes Quantum Volume (QV) to
determine the qubit circuit size for each device. Furthermore, we
utilize SupermarQ to determine each application’s benchmark
score, utilizing their 5 feature vectors to indicate the circuit
performance, and applying their circuit optimization techniques
to configure the transpiler. Our experiment shows that devicescore
of the simulated IBM backend are directly correlated to
Quantum Volume. However, it should be noted that topology
and gate error are also embedded in the benchmark scores of
the applications. Because of this, our study introduces the devicescore,
a single number metric that provides a more precise scale
with the assistance of Quantum Volume and SupermarQ.