Qubit Mapping is an essential step in realizing quantum circuits on actual hardware devices. However, due to the high complexity of this problem, current solutions can only work on circuits in fairly small scales (i.e. $<50$ qubits). In this paper, we propose a qubit mapping methodology which, to the best of our knowledge, is the first framework to handle very large quantum circuits (i.e. thousands of qubits) towards the quantum advantage. Our novel routing algorithm, Duostra, can efficiently identify the optimal routing path for a given two-qubit gate to operate on physical qubits through swap-gate insertions, and our scheduling heuristic offers the flexibility to strike the balance in optimizing the performance and pursuing the scalability. Experimental results show that our method runs $10$ times faster than the state-of-the-art approaches, while on average can still outperform them by over $5\%$ in terms of the execution time of the quantum circuits. More specifically, our proposed algorithm can complete the qubit mapping of an $11,969$-qubit Quantum Fourier Transform circuit within five hours.

翻译：量子比特映射是在实际硬件设备上实现量子电路的关键步骤。然而，由于该问题的高度复杂性，当前解决方案仅能处理较小规模的电路（即少于50个量子比特）。本文提出一种量子比特映射方法，据我们所知，这是首个能够处理超大规模量子电路（即数千个量子比特）以迈向量子优势的框架。我们新颖的路由算法Duostra能高效识别给定双量子比特门在物理量子比特上操作的最优路由路径（通过插入交换门实现），而我们的调度启发式策略提供了在优化性能与追求可扩展性之间取得平衡的灵活性。实验结果表明，我们的方法运行速度比现有最先进方法快10倍，同时在量子电路执行时间方面平均仍可优于它们超过5%。更具体地说，我们提出的算法能在五小时内完成对11,969量子比特量子傅里叶变换电路的量子比特映射。