Optimizing frequency allocation for superconducting quantum processors with frequency-tunable qubits

Abstract

As superconducting quantum processors scale, a key challenge is maintaining high coherence times and fidelity control over numerous qubits. We propose an automatic frequency allocation method for frequency-tunable qubits that equally considers coherence-limited fidelity and crosstalk-induced control errors during the allocation process. By employing a weighted average of the objective functions for coherence time and crosstalk, we numerically calculate gate fidelity to establish an open-loop optimization for determining suitable weight factors. This results in an efficient objective function for frequency optimization. We apply our method to frequency-tunable transmon qubits with tunable couplers, both theoretically and experimentally. The numerical results demonstrate significant advantages, including substantial reductions in gate errors and faster operation times, especially at higher qubit counts. Experimentally, our approach successfully achieves approximately 99.9% single-qubit fidelity on a nine-qubit chip.