Introducing qruise-toolset: a modular library for characterisation and creation of finely tailored digital twins for quantum computers

Abstract

Creating highly detailed and accurate digital twins of quantum computers is essential for understanding unexpected experimental results and for improving their designs and performance. Simulations that faithfully replicate the actual behaviour and dynamics of such systems are necessary for the fast prototyping of new QPU designs that overcome the shortcomings of their predecessors. We present qruise-toolset [1], a modular software stack that conveniently simulates the open quantum dynamics of noisy systems – seamlessly modelling both the QPU and the control stack. Applications include data-driven learning of system parameters using the differentiable digital twin [2], open loop optimisation for QPU operation, optimization of next-gen design and more [3, 4]. We discuss demonstrative applications such as characterisation and optimisation of system parameters for flux-based gates.

[1] https://www.qruise.com/products/qruise-ml

[2] J. Bradbury et al. http://github.com/jax-ml/jax

[3] AS. Roy et al. arXiv. 2205.04829 (2022).

[4] N. Wittler et al. Phys. Rev. Appl. 15.3 034080 (2021).

Authors:
Yousof Mardoukhi (Qruise GmbH)
Ana Gramajo (Qruise GmbH)
Satyanarayana Bade (Qruise GmbH)
Alastair Marshall (Qruise GmbH)
Anurag Saha Roy (Qruise GmbH)
Shai Machnes (Qruise GmbH)