arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2402.06210cs.AR
收藏

PULSE:面向低功耗稀疏感知卷积引擎的参数化硬件单元

PULSE: Parametric Hardware Units for Low-power Sparsity-Aware Convolution Engine

Ilkin Aliyev, Tosiron Adegbija

AI总结:

本文提出PULSE,一种参数化硬件生成器,根据层间稀疏性定制SNN加速器,通过优先级编码压缩和高效激活转移,在FPGA上实现3.14倍于传统加速器的效率提升。

摘要:

脉冲神经网络(SNN)因其比人工神经网络(ANN)更具生物真实性而变得流行。然而,在硬件中有效利用SNN固有的非结构化稀疏性具有挑战性,尤其是由于网络层间稀疏性的可变性。这种可变性取决于多种因素,包括输入数据集、编码方案和神经元模型。大多数现有的SNN加速器未能考虑应用(模型+数据集)的层特定工作负载,导致高能耗。为解决此问题,我们提出了一种设计时参数化硬件生成器,该生成器以层间稀疏性和处理单元数量为输入,并综合相应的硬件。所提出的设计使用优先级编码器压缩稀疏脉冲序列,并高效地在网络层间转移激活。我们通过首先分析给定应用的特征,然后执行高效的资源分配来展示所提方法的鲁棒性。在Xilinx Kintex FPGA(现场可编程门阵列)上使用MNIST、FashionMNIST和SVHN数据集的结果显示,与稀疏无感知的脉动阵列加速器相比,加速器效率(FPS/W)提高了3.14倍。与最近的稀疏感知工作相比,我们的解决方案效率提高了1.72倍。

英文摘要:

Spiking Neural Networks (SNNs) have become popular for their more bio-realistic behavior than Artificial Neural Networks (ANNs). However, effectively leveraging the intrinsic, unstructured sparsity of SNNs in hardware is challenging, especially due to the variability in sparsity across network layers. This variability depends on several factors, including the input dataset, encoding scheme, and neuron model. Most existing SNN accelerators fail to account for the layer-specific workloads of an application (model + dataset), leading to high energy consumption. To address this, we propose a design-time parametric hardware generator that takes layer-wise sparsity and the number of processing elements as inputs and synthesizes the corresponding hardware. The proposed design compresses sparse spike trains using a priority encoder and efficiently shifts the activations across the network's layers. We demonstrate the robustness of our proposed approach by first profiling a given application's characteristics followed by performing efficient resource allocation. Results on a Xilinx Kintex FPGA (Field Programmable Gate Arrays) using MNIST, FashionMNIST, and SVHN datasets show a 3.14x improvement in accelerator efficiency (FPS/W) compared to a sparsity-oblivious systolic array-based accelerator. Compared to the most recent sparsity-aware work, our solution improves efficiency by 1.72x.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑