Practical Quantum Computation via Circuit Partitioning and Qubit Reuse

Effective quantum computation on near-term hardware is, in part, constrained by a limited qubit count. In this work, we introduce a quantum circuit partitioning method that combines and builds upon existing techniques of circuit and gate cutting and qubit-reuse compilation. Our proposed approach can drastically reduce the number of qubits required for a quantum circuit by alternating steps of cutting and qubit reuse ordering.

Our method is tailored for quantum processors that allow mid-circuit measurements and resets, a feature that enables repurposing qubits during the computation. Unlike existing methods, our wire cutting approach can cut through the same wire multiple times and require up to a single ancilla qubit, leveraging mid-circuit measurement and reset capabilities. This reduces the number of required ancilla qubits by allowing their reuse. As a result, qubit reuse enables more ways to perform circuit partitioning.

We experimentally verify our work on the Quantinuum H1 trapped ion quantum processor, where we realize a specific class of ansatz circuits used in variational quantum algorithms that are well-suited for our partitioning technique. We demonstrate that, given a proper circuit implementation, our methods provide an evident improvement in physical qubit cost and the number of possible ways to perform circuit partitioning.