Fully automated comprehensive characterization and bring-up of superconducting quantum computers

Bringing up, characterizing and calibrating devices is a time-consuming, but essential step in experimental quantum computing. To tackle this problem, most laboratories still rely on home-brewed scripts that require significant manual intervention. While this approach may work for devices of modest sizes, it is inefficient and will not scale to processors with larger numbers of qubits.

In this talk, we introduce the Qruise scalable framework for automated bring-up and characterization of superconducting quantum processors. A user-defined directed acyclic graph of an atomic experiment is automatically traversed in a parallelized fashion. Atomic experiments can be either user-defined, or drawn from a library of over 30 distinct pre-defined experiments. The results at each stage are stored in a local database, and may be viewed through a web browser in real time. Concomitantly with the experiments, the Qruise framework also gradually builds up a detailed physical model of the device, along with an error budget designed to inform future experiments.

We showcase results of our bring-up framework and show how the resulting models are able to predict nontrivial device properties such as flux line transfer functions.