This repository includes a HLS-based template for the GNN-based JEDI-net with many hardware optimizaitons. These example designs are tested using Vivado HLS 2019.02 .
More can be found in our paper: https://arxiv.org/abs/2209.14065
We are still work on more examples which will be released later. If you find any issue, please ping me an email.
This is the design with the optimal accuracy and a low latency less than 1us.
- How to run:
cd jedi50p_opt_acc/prj_cmd01
vivado_hls -f build.tcl- How to run only C-simulation:
cd jedi50p_opt_acc/prj_cmd01
vivado_hls -f build_sim.tclThe reports are in the following directory: prj_cmd01/jedi_prj/solution1/syn/report/
Same to the steps above using a different directory.
jedi50p_opt_acc and jedi50p_opt_latn take us less than 1 hour to finish using a server with a Gold 6154 CPU. But jedi50p_baseline_u1 takes us 7 hours to finish the c-synthesis on the same server.
If you find our repository useful, please cite one of our papers:
Z. Que, M. Loo, H. Fan, M. Blott, M. Pierini, A. Tapper, W. Luk. "LL-GNN: Low Latency Graph Neural Networks on FPGAs for Particle Detectors", arXiv preprint arXiv:2209.14065, 2022.
PDF
Z. Que, M. Loo, H. Fan, M. Pierini, A. Tapper, W. Luk. "Optimizing graph Neural Networks for jet tagging in particle physics on FPGAs". In 2022 IEEE 32nd International Conference on Field Programmable Logic and Applications (FPL). IEEE, 2022.
PDF
Z. Que, M. Loo, W. Luk. "Reconfigurable Acceleration of Graph Neural Networks for Jet Identification in Particle Physics". In 2022 IEEE 4th International Conference on Artificial Intelligence Circuits and Systems (AICAS) (pp. 202-205). IEEE, 2022.
PDF