arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2311.04764cs.AR
收藏

AutoWS:自动化逐层流水线 DNN 加速器中的权重流式传输

AutoWS: Automate Weights Streaming in Layer-wise Pipelined DNN Accelerators

Zhewen Yu, Christos-Savvas Bouganis

AI总结:

本文提出 AutoWS 自动化存储管理方法,利用逐层流水线的静态调度协同片上与片外内存优化权重存储,支持大规模 DNN 在资源受限设备上部署。

摘要:

随着深度神经网络(DNN)取得巨大成功,高效硬件加速器的设计在研究界引发了广泛兴趣。现有研究探索了两种架构策略:顺序层执行和逐层流水线。前者支持更广泛的模型,而后者则因更强的定制化能力和更高的效率而受到青睐。逐层流水线架构面临的一个挑战是,其对用于权重存储的片上存储器有巨大需求,从而阻碍了大规模网络在资源受限设备上的部署。本文提出 AutoWS,这是一种开创性的存储器管理方法,利用逐层流水线架构的静态调度,同时使用片上和片外存储器来优化该架构中的权重存储。通过对硬件设计和设计空间探索进行全面研究,我们的方法实现了完全自动化,并使大规模 DNN 模型能够部署在资源受限设备上;而在现有面向逐层流水线架构的工作中,这是无法实现的。AutoWS 已开源:https://github.com/Yu-Zhewen/AutoWS

英文摘要:

With the great success of Deep Neural Networks (DNN), the design of efficient hardware accelerators has triggered wide interest in the research community. Existing research explores two architectural strategies: sequential layer execution and layer-wise pipelining. While the former supports a wider range of models, the latter is favoured for its enhanced customization and efficiency. A challenge for the layer-wise pipelining architecture is its substantial demand for the on-chip memory for weights storage, impeding the deployment of large-scale networks on resource-constrained devices. This paper introduces AutoWS, a pioneering memory management methodology that exploits both on-chip and off-chip memory to optimize weight storage within a layer-wise pipelining architecture, taking advantage of its static schedule. Through a comprehensive investigation on both the hardware design and the Design Space Exploration, our methodology is fully automated and enables the deployment of large-scale DNN models on resource-constrained devices, which was not possible in existing works that target layer-wise pipelining architectures. AutoWS is open-source: https://github.com/Yu-Zhewen/AutoWS

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。

↑