IEEE International Conference on Information Technology, Big Data and Artificial Intelligence (ICIBA)
High-Speed Parallel Predistortion Architecture for Z-Flux Bias Lines in Superconducting Quantum Computing
Maojie Ma, Yifei Chen, Liyu Tian, Qiongzhi Wu
Abstract
High-fidelity superconducting quantum computing is significantly limited by pulse waveform distortion in Z-flux bias lines. Implementing gigahertz real-time predistortion using traditional infinite impulse response (IIR) filters on FPGAs is challenging due to the feedback loop latency and numerical errors. To address this, we developed an efficient predistortion hardware architecture for high-speed arbitrary waveform generators. The architecture employed 8-point parallel data processing to unroll the IIR feedback calculation path, achieving stable operation at 2 GSa/s. We adopted a flexibly scalable multi-stage parallel structure to cover multi-exponential distortion combinations. Furthermore, we accurately extracted and deployed system parameters based on the constructed exponential distortion model and measured data. Results show that the architecture effectively compensates for Z-flux bias waveform distortion, accelerating computing speed by three orders of magnitude compared to hostside software execution. This provides an efficient hardware-level predistortion solution for large-scale superconducting quantum computing systems.