Speaker
Description
The High-Luminosity LHC (HL-LHC) will provide an order of magnitude increase in integrated luminosity and enhance the discovery reach for new phenomena. The increased pile-up foreseen during the HL-LHC necessitates major upgrades to the ATLAS detector and trigger. The Phase-II trigger will consist of two levels, a hardware-based Level-0 trigger and an Event Filter (EF) with tracking capabilities. Within the Trigger and Data Acquisition group, a heterogeneous computing farm consisting of CPUs and potentially GPUs and/or FPGAs is under study, together with the use of modern machine learning algorithms such as Graph Neural Networks (GNNs). GNNs are a powerful class of geometric deep learning methods for modeling spatial dependencies via message passing over graphs. They are well-suited for track reconstruction tasks by learning on an expressive structured graph representation of hit data and considerable speedup over CPU-based execution is possible on FPGAs. In this talk, we will provide novel results of fast track finding within a single FPGA using GNN for pattern recognition as part of the larger project for Phase-II EF system tracking on a FPGA. We present physics performance results and technical measurements related to power, throughput and latency aimed at motivating the viability of GNNs for pattern recognition as part of a fast tracking finding pipeline on an FPGA for the ATLAS Phase-II EF tracking system. We will highlight specific techniques employed to achieve our results, such as sequential processing of the detector, interaction network model hyperparameter tuning, model pruning and quantization-aware training.