Alibaba Cloud Quantum Development Platform: Surface Code Simulations with Crosstalk

Cupjin Huang, Xiaotong Ni, Fang Zhang, Michael Newman, Dawei Ding, Xun Gao, Tenghui Wang, Hui‐Hai Zhao, Feng Wu, Gengyan Zhang, Chunqing Deng, Hsiang‐Sheng Ku, Jianxin Chen, Yaoyun ShiarXiv 2020 · DOI · arXiv

Abstract

We report on our work using the Alibaba Cloud Quantum Development Platform (AC-QDP) to simulate a distance-3 logical qubit encoded in the 17-qubit surface code. Utilizing experimental noise parameters for transmon qubits in a planar circuit QED architecture, our simulation specifically accounts for crosstalk induced by ZZ-interactions. We demonstrate that at current-stage noise levels, crosstalk contributes significantly to the dephasing of the logical qubit, resulting in a total phase-flip probability of approximately 0.6%—roughly 60% higher than models excluding crosstalk. These findings indicate that while current hardware approaches the break-even fault-tolerance regime, crosstalk remains a critical barrier to achieving net-positive quantum error correction.

Introduction

Quantum Error Correction (QEC) is the cornerstone of scalable quantum computing, designed to protect fragile quantum information from environmental decoherence. The Surface Code is widely regarded as the most promising candidate for near-term implementation due to its high error threshold and local connectivity requirements. However, a significant gap persists between idealized theoretical models and the physical reality of superconducting hardware.

One of the most pervasive challenges in superconducting circuits is crosstalk, particularly unwanted ZZ-interactions where the state of one qubit shifts the frequency of its neighbor. While many QEC simulations rely on simplified stabilizer formalisms that ignore such physical nuances, this study utilizes the Alibaba Cloud Quantum Development Platform (AC-QDP) to quantify the impact of realistic crosstalk on the performance of the Surface-17 code.

System Architecture and Noise Modeling

The simulation focuses on the Surface-17 configuration, which consists of 9 data qubits and 8 ancilla qubits arranged in a planar lattice. This distance-3 code is the smallest surface code capable of correcting a single arbitrary error.

The Surface-17 code layout featuring 9 data qubits (blue squares) and 8 ancilla qubits (red and green circles).

To ensure experimental relevance, the model incorporates hardware-motivated parameters, including amplitude damping ($T_1 = 30\mu s$), phase damping ($T_\phi = 60\mu s$), gate-specific depolarizing noise, and readout errors. The primary focus is the ZZ-crosstalk strength ($k$), which is varied between 0.03 and 0.05 to observe its scaling effects on logical fidelity.

Simulation Methodology

The AC-QDP leverages the Tai-Zhang simulator, which employs Tensor Network Contraction rather than traditional state-vector methods. This approach offers two distinct advantages:

  1. Memory Efficiency: Tensor networks bypass the exponential memory requirements of storing full state vectors for many-qubit systems.
  2. Exact Noise Simulation: Unlike stochastic sampling (Monte Carlo) methods, tensor networks allow for the exact computation of logical Pauli Transition Matrices (PTMs). For high-precision requirements (e.g., $10^{-5}$ accuracy), this method is estimated to be nearly one million times faster than density-matrix sampling.

To incorporate continuous-time ZZ-crosstalk into a gate-based simulation, the authors discretized the interaction into instantaneous CPHASE gates. They rigorously validated this approximation, finding the "moving inaccuracy" to be negligible (on the order of $10^{-8}$), thereby justifying the discretization for large-scale simulations.

Syndrome extraction circuit illustrating the CZ region and the injection of idling and readout errors.

Results: The Impact of ZZ-Crosstalk

The most striking result of the simulation is the "Dephasing Tax" imposed by crosstalk. The study reveals that logical phase-flip errors are far more sensitive to ZZ-crosstalk than bit-flip errors. This asymmetry arises because X-type errors on ancilla qubits, exacerbated by photon decay, propagate through crosstalk channels to create "hook errors," which lead directly to logical phase-flips.

Key findings include:

  • Logical Error Overhead: With crosstalk included, the total phase-flip probability reaches ~0.6%. This represents a 60% increase in the error rate compared to a crosstalk-free model.
  • Linear Scaling: As crosstalk strength ($k$) increases toward 0.05, the phase-flip probability scales linearly, eventually exceeding 1%.
  • The Break-Even Threshold: The results suggest that current experimental noise levels are on the cusp of the break-even point—where a logical qubit's lifetime exceeds that of its physical constituents. However, the additional overhead from crosstalk currently keeps the system just below this critical threshold.

Conclusion and Future Work

This work highlights that crosstalk is not a marginal noise source but a primary driver of logical dephasing in the surface code. While the AC-QDP provides a powerful tool for exact noise simulation, several avenues for future research remain. Current simulations assume qubits remain within the computational subspace; however, "leakage" into higher energy states (qutrits) is a known issue in transmon hardware that requires more complex modeling. Future iterations of this research will incorporate leakage effects, active qubit reset schemes, and other crosstalk modalities (such as measurement-induced interference) to provide a comprehensive roadmap for hardware developers striving for fault tolerance.

Summary updated

Abstract

We report, in a sequence of notes, our work on the Alibaba Cloud Quantum Development Platform (AC-QDP). AC-QDP provides a set of tools for aiding the development of both quantum computing algorithms and quantum processors, and is powered by a large-scale classical simulator deployed on Alibaba Cloud. In this note, we simulate a distance-3 logical qubit encoded in the 17-qubit surface code using experimental noise parameters for transmon qubits in a planar circuit QED architecture. Our simulation features crosstalk induced by ZZ-interactions. We show that at the current-stage noise levels, crosstalk contributes significantly to the dephasing of the logical qubit. This results in a total phase-flip probability of $\sim 0.6\%$, about $60\%$ higher than expected without considering crosstalk. This indicates that for the code considered, the current noise parameters approach, but do not yet meet, the break-even fault-tolerance regime.
BibTeX
@misc{huang2020alibaba,
  author = {Cupjin Huang and Xiaotong Ni and Fang Zhang and Michael Newman and Dawei Ding and Xun Gao and Tenghui Wang and Hui‐Hai Zhao and Feng Wu and Gengyan Zhang and Chunqing Deng and Hsiang‐Sheng Ku and Jianxin Chen and Yaoyun Shi},
  title = {{Alibaba Cloud Quantum Development Platform: Surface Code Simulations with Crosstalk}},
  year = {2020},
  eprint = {2002.08918},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2002.08918},
}
Figure