Harnessing the Power of Long-Range Entanglement for Clifford Circuit Synthesis

TL;DR

This paper proposes methods for efficient synthesis of Clifford circuits in superconducting quantum architectures using long-range entanglement via GHZ states, achieving significant reductions in circuit complexity.

Contribution

It introduces entanglement-assisted models for Clifford circuit synthesis, providing complexity bounds and efficient schemes leveraging GHZ states and entanglement buses.

Findings

O(n^3) complexity for Clifford synthesis with one entanglement bus

Graph state synthesis with at most (n/2)+1 GHZ layers in a two-bus architecture

Clifford operations require about 1.5n GHZ layers in a two-bus setup

Abstract

In superconducting architectures, limited connectivity remains a significant challenge for the synthesis and compilation of quantum circuits. We consider models of entanglement-assisted computation where long-range operations are achieved through injections of large GHZ states. These are prepared using ancillary qubits acting as an ``entanglement bus,'' unlocking global operation primitives such as multi-qubit Pauli rotations and fan out gates. We derive bounds on the circuit size for several well studied problems, such as CZ circuit, CX circuit, and Clifford circuit synthesis. In particular, in an architecture using one such entanglement bus, we give an $O(n^3)$-complexity synthesis scheme for arbitrary Clifford operations requiring at most $2n + 1$ layers of entangled-state-injections. In a square-lattice architecture with two entanglement buses, we show that a graph state can be synthesized using at most $\lceil \frac{1}{2}n \rceil + 1$ layers of GHZ state injections, and Clifford operations require only $\lceil\frac{3}{2} n \rceil + O(\sqrt n)$ layers of GHZ state injections.