Building a virtual calibration through Dynamic Emulation on Superconducting Qubits scaling up and out via Accelerated Digital Twin

Jiun-I Lee^1*, Ana Laura Gramajo², Teik-Hui Lee¹

¹Research Center for Critical Issues, Tainan, Taiwan
²Qruise GmbH, Saarbrücken, Germany

* Presenter:Jiun-I Lee, email:as0200849@gate.sinica.edu.tw

Digital twin, which are essential in the era of artificial intelligence, could be utilized to create high-fidelity virtual replicas of quantum processors. This utilization could be critical to accelerate the emergence of fault tolerance quantum computers by reducing the 2q-error well below surface-code threshold. They provide a software mirror of physical qubit systems, allowing us to explore device behavior, understand noise impacts, and optimize control protocols without requiring constant use of the actual hardware. Such digital twin enable improved quantum gate optimization and tuning, support dynamic calibration procedures with feedback, and help refine error models for advanced noise mitigation strategies. In this work, we implement a digital twin of a superconducting quantum device by simulating its full quantum dynamics via the system Hamiltonian and the Lindblad master equation, capturing both coherent evolution and dissipative processes. The current focus is a SQUID-based transmon qubit model, which accurately represents a typical superconducting qubit including its anharmonic spectrum and noise channels. Notably, the modeling approach is general and can be extended to other qubit Hamiltonians and platforms beyond transmons. The framework is designed for scalable quantum systems, leveraging efficient numerical methods and high-performance computing resources. In particular, we employ GPU acceleration to dramatically improve simulation throughput, enabling the simulation of larger multiple-qubit systems This digital twin approach paves the way for more effective hardware-aware optimizations, guiding the development of quantum gates and error suppression techniques to enhance overall quantum processor performance. In doing so, it also lays critical groundwork for advancing artificial intelligence with quantum compute resources, moving us closer to an era where AI algorithms can directly harness QPU technology.

Keywords: Digital twin, Hamiltonian simulation, Superconducting quantum processors, GPU acceleration