Fast inference of deep neural networks in FPGAs for particle physics
Recent results at the Large Hadron Collider (LHC) have pointed to enhanced physics capabilities through the improvement of the real-time event processing techniques. Machine learning methods are ubiquitous and have proven to be very powerful in LHC physics, and particle physics as a whole. However,...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
IOP Publishing
2020
|
| Online Access: | https://hdl.handle.net/1721.1/128237 |
| _version_ | 1826200955760148480 |
|---|---|
| author | Han, Song Harris, Philip Coleman |
| author2 | Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science |
| author_facet | Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Han, Song Harris, Philip Coleman |
| author_sort | Han, Song |
| collection | MIT |
| description | Recent results at the Large Hadron Collider (LHC) have pointed to enhanced physics capabilities through the improvement of the real-time event processing techniques. Machine learning methods are ubiquitous and have proven to be very powerful in LHC physics, and particle physics as a whole. However, exploration of the use of such techniques in low-latency, low-power FPGA (Field Programmable Gate Array) hardware has only just begun. FPGA-based trigger and data acquisition systems have extremely low, sub-microsecond latency requirements that are unique to particle physics. We present a case study for neural network inference in FPGAs focusing on a classifier for jet substructure which would enable, among many other physics scenarios, searches for new dark sector particles and novel measurements of the Higgs boson. While we focus on a specific example, the lessons are far-reaching. A companion compiler package for this work is developed based on High-Level Synthesis (HLS) called hls4ml to build machine learning models in FPGAs. The use of HLS increases accessibility across a broad user community and allows for a drastic decrease in firmware development time. We map out FPGA resource usage and latency versus neural network hyperparameters to identify the problems in particle physics that would benefit from performing neural network inference with FPGAs. For our example jet substructure model, we fit well within the available resources of modern FPGAs with a latency on the scale of 100 ns. |
| first_indexed | 2024-09-23T11:44:21Z |
| format | Article |
| id | mit-1721.1/128237 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T11:44:21Z |
| publishDate | 2020 |
| publisher | IOP Publishing |
| record_format | dspace |
| spelling | mit-1721.1/1282372022-09-27T21:34:04Z Fast inference of deep neural networks in FPGAs for particle physics Han, Song Harris, Philip Coleman Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Massachusetts Institute of Technology. Department of Physics Recent results at the Large Hadron Collider (LHC) have pointed to enhanced physics capabilities through the improvement of the real-time event processing techniques. Machine learning methods are ubiquitous and have proven to be very powerful in LHC physics, and particle physics as a whole. However, exploration of the use of such techniques in low-latency, low-power FPGA (Field Programmable Gate Array) hardware has only just begun. FPGA-based trigger and data acquisition systems have extremely low, sub-microsecond latency requirements that are unique to particle physics. We present a case study for neural network inference in FPGAs focusing on a classifier for jet substructure which would enable, among many other physics scenarios, searches for new dark sector particles and novel measurements of the Higgs boson. While we focus on a specific example, the lessons are far-reaching. A companion compiler package for this work is developed based on High-Level Synthesis (HLS) called hls4ml to build machine learning models in FPGAs. The use of HLS increases accessibility across a broad user community and allows for a drastic decrease in firmware development time. We map out FPGA resource usage and latency versus neural network hyperparameters to identify the problems in particle physics that would benefit from performing neural network inference with FPGAs. For our example jet substructure model, we fit well within the available resources of modern FPGAs with a latency on the scale of 100 ns. 2020-10-29T14:37:24Z 2020-10-29T14:37:24Z 2018-07 2018-05 2020-10-27T15:02:23Z Article http://purl.org/eprint/type/JournalArticle 1748-0221 https://hdl.handle.net/1721.1/128237 Duarte, J. et al. “Fast inference of deep neural networks in FPGAs for particle physics.” Journal of Instrumentation, 13, 7 (July 2018) © 2018 The Author(s) en 10.1088/1748-0221/13/07/P07027 Journal of Instrumentation Creative Commons Attribution 3.0 unported license https://creativecommons.org/licenses/by/3.0/ application/pdf IOP Publishing IOP Publishing |
| spellingShingle | Han, Song Harris, Philip Coleman Fast inference of deep neural networks in FPGAs for particle physics |
| title | Fast inference of deep neural networks in FPGAs for particle physics |
| title_full | Fast inference of deep neural networks in FPGAs for particle physics |
| title_fullStr | Fast inference of deep neural networks in FPGAs for particle physics |
| title_full_unstemmed | Fast inference of deep neural networks in FPGAs for particle physics |
| title_short | Fast inference of deep neural networks in FPGAs for particle physics |
| title_sort | fast inference of deep neural networks in fpgas for particle physics |
| url | https://hdl.handle.net/1721.1/128237 |
| work_keys_str_mv | AT hansong fastinferenceofdeepneuralnetworksinfpgasforparticlephysics AT harrisphilipcoleman fastinferenceofdeepneuralnetworksinfpgasforparticlephysics |