Design of an FPGA-Based Neutral Atom Rearrangement Accelerator for Quantum Computing

TL;DR

This work proposes a novel quadrant-based rearrangement algorithm that employs a divide-and-conquer strategy and also enables the simultaneous movement of multiple atoms, even across different columns and rows, which makes this algorithm a promising solution for large-scale quantum systems.

Abstract

Neutral atoms have emerged as a promising technology for implementing quantum computers due to their scalability and long coherence times. However, the execution frequency of neutral atom quantum computers is constrained by image processing procedures, particularly the assembly of defect-free atom arrays, which is a crucial step in preparing qubits (atoms) for execution. To optimize this assembly process, we propose a novel quadrant-based rearrangement algorithm that employs a divide-and-conquer strategy and also enables the simultaneous movement of multiple atoms, even across different columns and rows. We implement the algorithm on FPGA to handle each quadrant independently (hardware-level optimization) while maximizing parallelization. To the best of our knowledge, this is the first hardware acceleration work for atom rearrangement, and it significantly reduces the processing time. This achievement also contributes to the ongoing efforts of tightly integrating quantum accelerators into High-Performance Computing (HPC) systems. Tested on a Zynq RFSoC FPGA at 250 MHz, our hardware implementation is able to complete the rearrangement process of a 30$\times$30 compact target array, derived from a 50$\times$50 initial loaded array, in approximately 1.0 $μs$. Compared to a comparable CPU implementation and to state-of-the-art FPGA work, we achieved about 54$\times$ and 300$\times$ speedups in the rearrangement analysis time, respectively. Additionally, the FPGA-based acceleration demonstrates good scalability, allowing for seamless adaptation to varying sizes of the atom array, which makes this algorithm a promising solution for large-scale quantum systems.

Figures (8)

• Figure 1: A typical workflow of neutral atom quantum computers. The image of the atom array is transformed into a binary representation, where black dots denote occupied areas, while white dots stand for positions where no atom is detected. This binary format serves as the input for the rearrangement algorithm. After making a schedule, the details of movements will be sent to the AWG, whose pulses control the AOD to tune the atoms.
• Figure 2: Architectures of atom control systems. (a) Current typical system structure: Detection and rearrangement processes are performed using CPUs or GPUs, where communications between components are needed. (b) Optimal (long-term) structure: All functional blocks are seamlessly integrated into , with detection and rearrangement implemented via customized hardware.
• Figure 3: Illustration of the typical rearrangement algorithm. The blue emphasized line indicates the target line for this move step. Arrows represent the atoms being moved in this step, with the direction of arrows indicating the direction of movement. The red square stands for the target filling area. Within one "Move" block, we can have multiple simultaneous moves. For instance, in "Move 1", there are empty holes in Rows 1, 3, 4, and 7, so we move all atoms positioned to the left of each hole, shifting them one step to the right.
• Figure 4: rearrangement schedule. By splitting the atom array and performing specific flip operations, we can apply a unified rearrangement method to each quadrant. This method provides an inherent acceleration on .
• Figure 5: Complete dataflow of the HLS-accelerated rearrangement module.