Shallower CNOT circuits on realistic quantum hardware

Timothée Goubault de Brugière,Simon Martiel

2023-03-14

Abstract:We focus on the depth optimization of CNOT circuits on hardwares with limited connectivity. We adapt the algorithm from Kutin et al. that implements any $n$-qubit CNOT circuit in depth at most $5n$ on a Linear Nearest Neighbour (LNN) architecture. Our proposal is a block version of Kutin et al.'s algorithm that is scalable with the number of interactions available in the hardware: the more interactions we have the less the depth. We derive better theoretical upper bounds and we provide a simple implementation of the algorithm. Overall, we achieve better depth complexity for CNOT circuits on some realistic quantum hardware like a grid or a ladder. For instance the execution of a $n$-qubit CNOT circuit on a grid can be done in depth $4n$.

Quantum Physics

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to optimize the depth of CNOT circuits on actual quantum hardware with limited connectivity. Specifically, the paper focuses on how to achieve a shallower CNOT circuit depth under hardware constraints, such as the Linear Nearest Neighbor (LNN) architecture. The CNOT gate is a two - qubit gate. If two qubits are too far apart in the hardware, their interaction cannot be physically realized, and thus the CNOT gate cannot be executed. This limits the qubit pairs to which the CNOT gate can be applied and adds additional constraints to the compiler.

The main contribution of the paper is to propose a block version of the Kutin et al. algorithm, which is applicable to any architecture that can group qubits into groups of the same size and these groups are arranged in a line. For example, in a trapezoidal architecture such as IBM QX5, groups of size 2 can be used; in a grid architecture, groups of size 4 can be used. Although this algorithm is not a general solution for arbitrary qubit connectivity, it is flexible enough to be applicable to realistic quantum architectures.

Through this method, the authors show that on some actual quantum hardware, such as grid or trapezoidal architectures, better CNOT circuit depth complexity can be achieved. For example, executing an n - qubit CNOT circuit on a grid architecture can be completed within a depth of 4n.