An Optical XNOR-Bitcount Based Accelerator for Efficient Inference of Binary Neural Networks
An Optical XNOR-Bitcount Based Accelerator for Efficient Inference of Binary Neural Networks
复制标题
DOI：
10.1109/isqed57927.2023.10129294
复制
发表时间：
2023-02
期刊：
影响因子：
--
通讯作者：
Sairam Sri Vatsavai;Venkata Sai Praneeth Karempudi;Ishan G. Thakkar
中科院分区：
文献类型：
--
作者：
Sairam Sri Vatsavai;Venkata Sai Praneeth Karempudi;Ishan G. Thakkar
与全精度卷积神经网络(CNN)相比，二进制神经网络(BNN)越来越受到人们的青睐，以最小的精度损失减少推理处理的内存和计算需求。BNN将CNN模型参数转换为1位精度，允许通过简单的XNOR和位计数操作处理BNN的推断。这使得BNN服从于硬件加速。已经提出了几种基于光子集成电路(PIC)的BNN加速器。虽然这些加速器提供的吞吐量和能效明显高于它们的电子同行，但这些加速器中利用的XNOR和位计数电路需要进一步增强，以提高它们的面积、能效和吞吐量。本文旨在满足这一需求。为此，我们发明了一种基于单MRR的光学异或门(OXG)。此外，我们还提出了一种新的位计数电路设计，称为光电荷累加器(PCA)。我们利用密集波分复用(DWDM)将多个OXG级联，并将它们连接到PCA，从而构建了一种新型的基于光学XNOR-Bitcount的二进制神经网络加速器(OXBNN)。我们对四个现代BNN的评估表明，OXBNN在帧/秒(FPS)和FPS/W(能量效率)方面分别比两个基于PIC的BNN加速器的几何平均值提高了62×和7.6×。我们开发了一个基于事务级别、事件驱动的基于Python的加速器评估模拟器(https://github.com/uky-UCAT/B_ONN_SIM).
Binary Neural Networks (BNNs) are increasingly preferred over full-precision Convolutional Neural Networks (CNNs) to reduce the memory and computational requirements of inference processing with minimal accuracy drop. BNNs convert CNN model parameters to 1-bit precision, allowing inference of BNNs to be processed with simple XNOR and bitcount operations. This makes BNNs amenable to hardware acceleration. Several photonic integrated circuits (PICs) based BNN accelerators have been proposed. Although these accelerators provide remarkably higher throughput and energy efficiency than their electronic counterparts, the utilized XNOR and bitcount circuits in these accelerators need to be further enhanced to improve their area, energy efficiency, and throughput. This paper aims to fulfill this need. For that, we invent a single-MRR-based optical XNOR gate (OXG). Moreover, we present a novel design of bitcount circuit which we refer to as Photo-Charge Accumulator (PCA). We employ multiple OXGs in a cascaded manner using dense wavelength division multiplexing (DWDM) and connect them to the PCA, to forge a novel Optical XNOR-Bitcount based Binary Neural Network Accelerator (OXBNN). Our evaluation for the inference of four modern BNNs indicates that OXBNN provides improvements of up to 62× and 7.6× in frames-persecond (FPS) and FPS/W (energy efficiency), respectively, on geometric mean over two PIC-based BNN accelerators from prior work. We developed a transaction-level, event-driven pythonbased simulator for evaluation of accelerators (https://github.com/uky-UCAT/B_ONN_SIM).