Reconfigurable Computing Challenge: Real-Time Graph Neural Networks for Online Event Selection in Big Science

TL;DR

This paper demonstrates a real-time, FPGA-accelerated graph neural network system for collider experiment triggers, achieving high throughput and low latency with a semi-automated design flow.

Contribution

It introduces an end-to-end demonstrator deploying a dynamic GNN on FPGA and AI Engine tiles, with a novel Python-based design flow for optimization.

Findings

Achieved 2.94 million events/sec throughput with 7.15 μs latency.

53% throughput improvement over FPGA-only baseline.

Reduced DSP utilization from 99% to 19%.

Abstract

Graph neural networks are increasingly adopted in trigger systems for collider experiments, where strict latency and throughput constraints render deployment on embedded platforms challenging. As detectors move towards higher granularity, the number of inputs per inference increase and FPGA-only solutions face resource bottlenecks. This work presents an end-to-end demonstrator for the real-time deployment of a dynamic Graph Neural Network for the Belle II electromagnetic calorimeter hardware trigger on the AMD Versal VCK190, leveraging both FPGA fabric and AI Engine tiles. We develop a Python-based semi-automated design flow covering operator fusion, partitioning, mapping, spatial parallelization, and kernel-level optimization. Our design achieves a throughput of 2.94 million events per second at an end-to-end latency of 7.15 microseconds. Compared to the FPGA-only baseline, this represents a 53% throughput improvement while reducing DSP utilization from 99% to 19% at 29% AI Engine tile utilization. To validate the deployment, an interactive visualization pipeline enables real-time monitoring of inference results on the physical demonstrator.