MultiQ: Multi-Programming Neutral Atom Quantum Architectures

TL;DR

This work addresses the throughput bottleneck in neutral-atom QPUs caused by large fidelity drop for big circuits and substantial initialization latency for small ones. It introduces MultiQ, a compiler-controller-checker co-design that enables co-execution of multiple circuits on a single NA QPU by generating virtual zone layouts, bundling and parallelizing circuits, and verifying functional independence with ZX-calculus-based techniques. The system comprises a compiler that emits virtual layouts and QASM^mq IR, a runtime controller that bundles and places circuits to maximize spatial and temporal QPU utilization, and a checker that guarantees correctness of multi-programmed executions. On 11 benchmarks, MultiQ delivers throughput gains from 3.8x to 12.3x with minimal fidelity loss (1.3% to 3.5%), demonstrating substantial improvements in hardware utilization and reduced initialization overhead, while preserving the semantics of each circuit. The work thus provides a scalable, cross-layer solution for high-throughput, fidelity-conscious multi-programming on neutral-atom architectures and includes open-source artifacts for broad adoption.

Abstract

Neutral atom Quantum Processing Units (QPUs) are emerging as a popular quantum computing technology due to their large qubit counts and flexible connectivity. However, performance challenges arise as large circuits experience significant fidelity drops, while small circuits underutilize hardware and face initialization latency issues. To tackle these problems, we propose $\textit{multi-programming on neutral atom QPUs}$, allowing the co-execution of multiple circuits by logically partitioning the qubit array. This approach increases resource utilization and mitigates initialization latency while maintaining result fidelity. Currently, state-of-the-art compilers for neutral atom architectures do not support multi-programming. To fill this gap, we introduce MultiQ, the first system designed for this purpose. MultiQ addresses three main challenges: (i) it compiles circuits into a $\textit{virtual zone layout}$ to optimize spatio-temporal hardware utilization; (ii) it parallelizes the execution of co-located circuits, allowing single hardware instructions to operate on different circuits; and (iii) it includes an algorithm to verify the functional independence of the bundled circuits. MultiQ functions as a cross-layer system comprising a compiler, controller, and checker. Our compiler generates \emph{virtual zone layouts} to enhance performance, while the controller efficiently maps these layouts onto the hardware and resolves any conflicts. The checker ensures the correct bundling of circuits. Experimental results show a throughput increase from 3.8$\times$ to 12.3$\times$ when multi-programming 4 to 14 circuits, with fidelity largely maintained, ranging from a 1.3% improvement for four circuits to only a 3.5% loss for fourteen circuits. Overall, MultiQ facilitates concurrent execution of multiple quantum circuits, boosting throughput and hardware utilization.

Figures (19)

• Figure 1: (a) Limitations of neutral atom QPUs evaluated using state-of-the-art NA compilers (ZAC lin2024reuseawarecompilationzonedquantum, PachinQo ludmir_modeling_2024 and Atomique wang_atomique_2024). (a) Fidelity drops drastically with circuit size, leading to QPU underutilization. (b) Circuit execution time is lower compared to QPU initialization time for circuits up to 170 qubits.
• Figure 2: Neutral atoms architecture basics (§ \ref{['section:background']}) Storage, entanglement, and measurement zones distributions and their standard atom and zone spacings. Single-qubit gates and two-qubit gate operations. AOD laser targeting and the non-overlapping constraints.
• Figure 3: (a) Tradeoff between QPU utilization and circuit shuttling time. Narrow layouts (orange) fully utilize the QPU but incur long shuttling operations. Wider layouts (red) minimize shuttling times but incur low utilization. Balanced layout (green).(b) Circuit bundling. Bundling circuits from the input queue (top) into execution bins (bottom) involves finding a solution that maximizes both spatial and temporal QPU utilization. Here, Option #2 reduces the total execution runtime.
• Figure 4: (a) Relative fidelity of two layouts compared to the square layout (ratio 1:1), with increasing circuit size. Narrow layouts (blue bars, 1:4 ratio) achieve lower fidelity than the square ones, while wide layouts (orange bars, 4:1 ratio) achieve higher. (b) Total shuttling time with increasing QPU utilization for ZAC lin2024reuseawarecompilationzonedquantum, executing circuit sequentially (single), circuits merged in parallel (grouped), and concurrently and independently (grouped independent)
• Figure 5: Overview of MultiQ (§ \ref{['sec:overview']}). MultiQ is a co-designed compiler-controller system. The compiler (§ \ref{['sec:compiler']}) optimizes and compiles quantum circuits. The controller (§ \ref{['sec:controller']}) then maps and efficiently multiprograms them on the hardware. A functional independence checker (§ \ref{['sec:checker']}) verifies that the instructions maintain circuit functionality.