Cargando…

Generalized Machine Learning Quantization Implementation for High Level Synthesis Targeting FPGAs

The Large Hadron Collider produces a large amount of data while operating, approximately one petabyte of data per second. The collider is currently undergoing an upgrade to collide more particles and produce even more data. In order to handle this large quantity of data, high throughput and low late...

Descripción completa

Detalles Bibliográficos
Autor principal: Trahms, Matthew Karl
Lenguaje:eng
Publicado: 2022
Materias:
Acceso en línea:http://cds.cern.ch/record/2804953
_version_ 1780972901163335680
author Trahms, Matthew Karl
author_facet Trahms, Matthew Karl
author_sort Trahms, Matthew Karl
collection CERN
description The Large Hadron Collider produces a large amount of data while operating, approximately one petabyte of data per second. The collider is currently undergoing an upgrade to collide more particles and produce even more data. In order to handle this large quantity of data, high throughput and low latency algorithms are required to filter interesting collision results out of the rest of the data collected by the sensors attached to the collider. Machine learning algorithms can be used for this filtering task with comparable accuracy to the traditional filtering algorithms and provide a wide range of accelerator options. FINN and hls4ml are frameworks to deploy machine learning models on Field Programmable Gate Arrays for high throughput, low latency acceleration options. FINN utilizes Brevitas, a quantization aware training library. Using Brevitas, I trained a particle tracking network and demonstrated equivalent accuracy at lower bit precision than post training quantization. As a cross organizational project, the hls4ml and FINN teams collaborated to develop the QONNX standard for quantized machine learning model representation. In order to integrate QONNX into hls4ml, I implemented new transformations to support the unique structures of QONNX.
id cern-2804953
institution Organización Europea para la Investigación Nuclear
language eng
publishDate 2022
record_format invenio
spelling cern-28049532022-03-29T20:41:10Zhttp://cds.cern.ch/record/2804953engTrahms, Matthew KarlGeneralized Machine Learning Quantization Implementation for High Level Synthesis Targeting FPGAsComputing and ComputersDetectors and Experimental TechniquesThe Large Hadron Collider produces a large amount of data while operating, approximately one petabyte of data per second. The collider is currently undergoing an upgrade to collide more particles and produce even more data. In order to handle this large quantity of data, high throughput and low latency algorithms are required to filter interesting collision results out of the rest of the data collected by the sensors attached to the collider. Machine learning algorithms can be used for this filtering task with comparable accuracy to the traditional filtering algorithms and provide a wide range of accelerator options. FINN and hls4ml are frameworks to deploy machine learning models on Field Programmable Gate Arrays for high throughput, low latency acceleration options. FINN utilizes Brevitas, a quantization aware training library. Using Brevitas, I trained a particle tracking network and demonstrated equivalent accuracy at lower bit precision than post training quantization. As a cross organizational project, the hls4ml and FINN teams collaborated to develop the QONNX standard for quantized machine learning model representation. In order to integrate QONNX into hls4ml, I implemented new transformations to support the unique structures of QONNX.CERN-THESIS-2022-019oai:cds.cern.ch:28049532022-03-25T17:38:11Z
spellingShingle Computing and Computers
Detectors and Experimental Techniques
Trahms, Matthew Karl
Generalized Machine Learning Quantization Implementation for High Level Synthesis Targeting FPGAs
title Generalized Machine Learning Quantization Implementation for High Level Synthesis Targeting FPGAs
title_full Generalized Machine Learning Quantization Implementation for High Level Synthesis Targeting FPGAs
title_fullStr Generalized Machine Learning Quantization Implementation for High Level Synthesis Targeting FPGAs
title_full_unstemmed Generalized Machine Learning Quantization Implementation for High Level Synthesis Targeting FPGAs
title_short Generalized Machine Learning Quantization Implementation for High Level Synthesis Targeting FPGAs
title_sort generalized machine learning quantization implementation for high level synthesis targeting fpgas
topic Computing and Computers
Detectors and Experimental Techniques
url http://cds.cern.ch/record/2804953
work_keys_str_mv AT trahmsmatthewkarl generalizedmachinelearningquantizationimplementationforhighlevelsynthesistargetingfpgas