Low Latency Edge Classification GNN for Particle Trajectory Tracking on FPGAs
This work addresses low latency particle tracking for high-energy physics experiments, representing an incremental improvement in optimizing GNNs for specific hardware constraints.
The paper tackled the challenge of in-time particle trajectory reconstruction in the Large Hadron Collider by introducing a resource-efficient GNN architecture on FPGAs, achieving 1625x and 1574x performance improvements over CPU and GPU respectively.
In-time particle trajectory reconstruction in the Large Hadron Collider is challenging due to the high collision rate and numerous particle hits. Using GNN (Graph Neural Network) on FPGA has enabled superior accuracy with flexible trajectory classification. However, existing GNN architectures have inefficient resource usage and insufficient parallelism for edge classification. This paper introduces a resource-efficient GNN architecture on FPGAs for low latency particle tracking. The modular architecture facilitates design scalability to support large graphs. Leveraging the geometric properties of hit detectors further reduces graph complexity and resource usage. Our results on Xilinx UltraScale+ VU9P demonstrate 1625x and 1574x performance improvement over CPU and GPU respectively.