ZAP: Zoned Architecture and Parallelizable Compiler for Field Programmable Atom Array

TL;DR

This paper introduces ZAP, a zoned architecture and compiler for neutral atom quantum computers that improves scalability, parallelism, and fidelity by optimizing atom placement and interaction scheduling.

Contribution

The paper presents a novel zoned architecture and a tailored compiler algorithm that significantly enhance quantum compilation efficiency and fidelity for neutral atom systems.

Findings

Achieves 5.4x increase in fidelity for 100-qubit systems.

Reduces operation depth and enhances parallelism in quantum compilation.

Demonstrates scalability and efficiency improvements over existing methods.

Abstract

Neutral atom quantum computing platforms have gained significant attention due to their potential scalability and flexibility in qubit arrangement. In this work, we present a novel zoned architecture for neutral atom quantum compilation, which divides the system into distinct zones: a storage zone and an interaction zone. This architecture optimizes atom placement and interaction scheduling, effectively reducing the operation depth and improving parallelism during compilation. Through a tailored algorithmic approach, we significantly enhance the compilation efficiency and scalability compared to existing methods. Compared to the state-of-the-art Enola platform, our method achieves a 5.4x increase in fidelity when the system need 100 qubits, marking a pivotal advancement in neutral atom quantum computing. Our approach provides a robust framework for future large-scale quantum computations, ensuring both high fidelity and efficient execution.