AtomFlow End-to-End FPGA Based Control Architecture for Neutral Atom Quantum Computers

Jul 16, 2026

Upload by Aurelia

Neutral Atom Quantum Computing (NAQC) promises scalable quantum processors thanks to long coherence times and naturally identical atomic qubits, yet suffers from slow execution caused by classical bottlenecks like fluorescence imaging, cooling, and atom rearrangement. AtomFlow introduces an end-to-end FPGA-based control architecture that merges fluorescence image analysis with a new atom-rearrangement algorithm onto a single Zynq UltraScale+ device. Co-locating both stages enables streaming rearrangement moves, removing host round-trip latency. On a 16×16 atom array, AtomFlow reaches 25.3 ms end-to-end latency with 4 ms first-move latency and 1 ms average move generation.

LanguageEnglish

CategoryTechnology

FormatPDF

Pages10

Reading Time25 min

Introduction

Background: latency bottlenecks in neutral atom quantum control

Motivation: software optimization vs hardware acceleration

Current control schematic and pipeline stages

arXiv :2607 . 11490v1 [ quant-ph] 13 Jul 2026 AtomFlow: An End-to-End FPGA-Based Control Architecture for Neutral Atom Quantum Computers Xiaorang Guo∗ , Jonas Winklmann∗ , Vengkeat Chea and Martin Schulz Chair of Computer Architecture and Parallel Systems TUM School of Computation, Information and Technology (CIT) Technical University of Munich, Garching, Germany Email: {xiaorang.guo, jonas.winklmann, vengkeat.chea, [martin.w.j.schulz](martin.w.j.schulz}@tum.de)[}](martin.w.j.schulz}@tum.de)[@tum.de](martin.w.j.schulz}@tum.de) Abstract—Neutral Atom Quantum Computing (NAQC) is an emerging modality for scalable quantum computation, valued for its long coherence times and the naturally identical atomic qubits. However, one of the main drawbacks is its slow execution rate, dominated by lengthy classical processing tasks, such as fluorescence imaging, cooling, and atom rearrangement. We address this bottleneck with AtomFlow, a field-programmable gate array (FPGA)-based control architecture that consolidates fluorescenceimage analysis and a newly developed atom-rearrangement algorithm onto a single Zynq UltraScale+ device. By co-locating the two stages on the same board and emitting rearrangement moves in a streaming fashion as soon as they are computed, AtomFlow eliminates the round-trip latency of conventional host-mediated pipelines. Evaluated on a 16×16 atom array, AtomFlow achievesan end-to-end latency of 25.3 ms with a first-move latency of 4 ms and an average move generation of 1 ms. Furthermore, our scalability analysis demonstrates that the architecture can readily support larger atom arrays within a single-board resource budget. Index Terms—Quantum Computing, Neutral Atoms, FPGA, Atom Detection, Rearrangement I. INTRODUCTION Achieving practical quantum advantage has become a central task in quantum computing [1], driving rapid progress across physical qubit modalities such as superconducting qubits [2], trapped ions [3], photonic systems [4], and neutral atoms (NAs) [5], [6] . Among these, NAs are increasingly recognized as a leading candidate for large-scale quantum processors, mainly due to their long coherence times [7], goodscalability [8], and flexible connectivity between atoms [9] . However, within the control pipeline of neutral atom quantum computers (NAQCs), the initialization and readout stages remain dominant latency bottlenecks. These stages are particularly critical when executing complex quantum programs under limited coherence-time budgets, and become even more demanding in the presence of quantum error correction (QEC), which requires repeated rounds of high-fidelity readout. ∗ Equal Contribution. This work was funded by the German Federal Ministry of Research, Technology and Space (BMFTR) under the funding program Quantum TechnologiesFrom Basic Research to Market under contract number 13N16087, as well as from the Munich Quantum Valley (MQV), which is supported by the Bavarian State Government with funds from the Hightech Agenda Bayern. Fig. 1: Abstract Representation of the current control schematic of NAQCs. In NAQCs, qubit states are encoded in quantum states of individual atoms and manipulated via laser-driven operations [10] . Prior to gate execution, atoms are loaded probabilistically into a two-dimensional array of optical traps, resulting in random occupancy states that complicate circuit mapping. Consequently, an initialization procedure is required, consisting of atom detection through fluorescence imaging, followed by rearrangement of present atoms to construct a smaller defect-free array. During program execution, measurements are also performed, both to produce final outputs and to enable conditional operations through mid-circuit feedback. In both cases, a reliable image processing mechanism is indispensable. Moreover, image processing and rearrangement are highly algorithmically intensive stages in the control pipeline, and their efficiency directly determines sy

FAQs

What problem does AtomFlow address in neutral atom quantum computing?

AtomFlow targets the slow execution caused by classical processing bottlenecks, especially fluorescence imaging and atom rearrangement, which dominate overall latency.

How does AtomFlow reduce latency compared with host-mediated pipelines?

AtomFlow co-locates fluorescence image analysis and atom-rearrangement computation on a single FPGA device and streams rearrangement moves immediately after they are computed, avoiding host-backend round-trip latency.

What performance does AtomFlow achieve on a 16×16 atom array?

AtomFlow reports an end-to-end latency of 25.3 ms, a first-move latency of 4 ms, and an average move generation time of 1 ms.

Jul 16, 2026

Upload by Aurelia