A-GHOST: High-rate streaming of trigger-level data to programmable GPU inference
Abstract
A-GHOST (A Global Heterogeneous Online Scouting Trigger) is an R&D effort investigating high-rate streaming readout architectures for high-energy physics (HEP) experiments. The central idea is to convert the deterministic front-end and fixed-latency processing close to the detector from a decision-making platform to an aggregation and streaming source. Compact trigger-level data are streamed to a GPU-enabled backend, where substantially more complex algorithms can run beyond the resource and latency envelope of the hardware trigger. This paper presents a proof-of-concept backend using the NVIDIA IGX Thor development kit. A software rate-controlled transmitter sends data through a QSFP interface and external loopback cable to a second QSFP interface, allowing the network-to-GPU path to be studied before integration with FPGA sources. NVIDIA DAQIRI enables reception directly into GPU-accessible memory, while a custom CUDA kernel reassembles packet payloads into persistent, contiguous TensorRT input windows without data-type conversion, which is handled by the models. Using a HEP-derived event representation consisting of the ten leading calorimeter clusters, the backend scales from 40 to 100 Gbps while sustaining a constant input rate and an inference latency of 0.118 ms at the p95 interval. The system operates for multiple hours without drops or inference errors. Generative and temporally aware neural networks demonstrate workloads beyond those normally deployed in fixed-latency FPGA trigger systems. The result establishes A-GHOST as a basis for future FPGA-to-GPU streaming readout systems.
Related
- Performance of heavy-flavour jet identification in the CMS high-level trigger in proton-proton collisions at √s = 13.6 TeV
- Complete set of tree-level 2→ 2 scattering amplitudes of ghost-free bimetric theory
- SCRAMJet: An Ultra Low Latency JetClass-Scale Transformer for the HL-LHC Level-1 Trigger
- Knowledge Distillation of a Normalising Flow for Real-Time Anomaly Detection at LHC Level-1 Trigger
- Hypervectors from Optical Disorder: Programmable Encoding and Optical Inference for Hyperdimensional Computing
BibTeX
@misc{xiotidis2026aghost,
title = {{A-GHOST: High-rate streaming of trigger-level data to programmable GPU inference}},
author = {I. Xiotidis and N. Clarke Hall and M. S. Larson and R. Gurunathan and C. Burdick and T. Du and N. Konstantinidis and K. Kordas and D. Leshchev and D. W. Miller and V. A. Petrovic and D. Sampsonidou and A. Thompson and T. Wengler},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.01761},
eprint = {2610.01761},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.01761},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - A-GHOST: High-rate streaming of trigger-level data to programmable GPU inference
AU - I. Xiotidis
AU - N. Clarke Hall
AU - M. S. Larson
AU - R. Gurunathan
AU - C. Burdick
AU - T. Du
AU - N. Konstantinidis
AU - K. Kordas
AU - D. Leshchev
AU - D. W. Miller
AU - V. A. Petrovic
AU - D. Sampsonidou
AU - A. Thompson
AU - T. Wengler
PY - 2026
DA - 2026/10/01
PB - arXiv
DO - 10.48550/arxiv.2610.01761
UR - https://arxiv.org/abs/2610.01761
AB - A-GHOST (A Global Heterogeneous Online Scouting Trigger) is an R&D effort investigating high-rate streaming readout architectures for high-energy physics (HEP) experiments. The central idea is to convert the deterministic front-end and fixed-latency processing close to the detector from a decision-making platform to an aggregation and streaming source. Compact trigger-level data are streamed to a GPU-enabled backend, where substantially more complex algorithms can run beyond the resource and latency envelope of the hardware trigger. This paper presents a proof-of-concept backend using the NVIDIA IGX Thor development kit. A software rate-controlled transmitter sends data through a QSFP interface and external loopback cable to a second QSFP interface, allowing the network-to-GPU path to be studied before integration with FPGA sources. NVIDIA DAQIRI enables reception directly into GPU-accessible memory, while a custom CUDA kernel reassembles packet payloads into persistent, contiguous TensorRT input windows without data-type conversion, which is handled by the models. Using a HEP-derived event representation consisting of the ten leading calorimeter clusters, the backend scales from 40 to 100 Gbps while sustaining a constant input rate and an inference latency of 0.118 ms at the p95 interval. The system operates for multiple hours without drops or inference errors. Generative and temporally aware neural networks demonstrate workloads beyond those normally deployed in fixed-latency FPGA trigger systems. The result establishes A-GHOST as a basis for future FPGA-to-GPU streaming readout systems.
ER -