PolyLUT-Add: FPGA-based LUT Inference with Wide Inputs

Autor:	Lou, Binglei, Rademacher, Richard, Boland, David, Leong, Philip H. W.
Rok vydání:	2024
Předmět:	Computer Science - Machine Learning Computer Science - Artificial Intelligence Computer Science - Hardware Architecture
Zdroj:	International Conference on Field-Programmable Logic and Applications (FPL2024) in Turin, Italy, from 2nd to 6th September 2024
Druh dokumentu:	Working Paper
Popis:	FPGAs have distinct advantages as a technology for deploying deep neural networks (DNNs) at the edge. Lookup Table (LUT) based networks, where neurons are directly modeled using LUTs, help maximize this promise of offering ultra-low latency and high area efficiency on FPGAs. Unfortunately, LUT resource usage scales exponentially with the number of inputs to the LUT, restricting PolyLUT to small LUT sizes. This work introduces PolyLUT-Add, a technique that enhances neuron connectivity by combining $A$ PolyLUT sub-neurons via addition to improve accuracy. Moreover, we describe a novel architecture to improve its scalability. We evaluated our implementation over the MNIST, Jet Substructure classification, and Network Intrusion Detection benchmark and found that for similar accuracy, PolyLUT-Add achieves a LUT reduction of $2.0-13.9\times$ with a $1.2-1.6\times$ decrease in latency. Comment: The source code for this paper is available at: https://github.com/bingleilou/PolyLUT-Add
Databáze:	arXiv
Externí odkaz:	http://arxiv.org/abs/2406.04910 Zobrazit plný text záznamu View this record from Arxiv