当前位置: X-MOL 学术IEEE J. Emerg. Sel. Top. Circuits Syst. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
SimuNN: A Pre-RTL Inference, Simulation and Evaluation Framework for Neural Networks
IEEE Journal on Emerging and Selected Topics in Circuits and Systems ( IF 3.7 ) Pub Date : 2020-06-01 , DOI: 10.1109/jetcas.2020.2993854
Shan Cao , Wei Deng , Zhenyi Bao , Chengbo Xue , Shugong Xu , Shunqing Zhang

Neural networks have been widely deployed in a number of applications due to their strong learning and feature extraction ability. To meet the ever increasing accuracy requirements from various applications, neural network models have become more complicated and diversified by exploiting more layers, larger model size, and more diverse functions. As a result, the design of application specified hardware accelerators for neural networks is also becoming more difficult, especially for real-time embedded applications. To efficiently bridge the gap between algorithm and hardware design phases, a pre-RTL neural network simulator (SimuNN) is proposed in this paper to enable early phase verification and fast prototyping. SimuNN can be used for inference, simulation and also evaluation, results of which can guide the design of both neural network models and hardware accelerators. SimuNN supports inference in various data precision, and is compatible with TensorFlow, which makes it easy for platform migration. It provides multi-level trace results that can be taken as the gold model of RTL designs. Moreover, the hardware performance under various quantizations, dataflow organizations and hardware configurations can be evaluated by SimuNN based on a generalized hardware model. Based on that, two dataflow organization schemes are concluded for determining the optimal configurations of hardware architecture under various hardware and performance constraints.

中文翻译:

SimuNN:神经网络的预 RTL 推理、模拟和评估框架

神经网络由于其强大的学习和特征提取能力而被广泛部署在许多应用中。为了满足各种应用不断提高的精度要求,神经网络模型通过开发更多的层、更大的模型尺寸和更多样的功能而变得更加复杂和多样化。因此,为神经网络设计应用程序指定的硬件加速器也变得越来越困难,尤其是对于实时嵌入式应用程序。为了有效地弥合算法和硬件设计阶段之间的差距,本文提出了一种预 RTL 神经网络模拟器 (SimuNN),以实现早期阶段验证和快速原型设计。SimuNN 可用于推理、模拟和评估,其结果可以指导神经网络模型和硬件加速器的设计。SimuNN 支持各种数据精度的推理,并兼容 TensorFlow,便于平台迁移。它提供了多级跟踪结果,可作为 RTL 设计的黄金模型。此外,基于广义硬件模型的 SimuNN 可以评估各种量化、数据流组织和硬件配置下的硬件性能。在此基础上,总结出两种数据流组织方案,用于在各种硬件和性能约束下确定硬件架构的最佳配置。它提供了多级跟踪结果,可作为 RTL 设计的黄金模型。此外,基于广义硬件模型的 SimuNN 可以评估各种量化、数据流组织和硬件配置下的硬件性能。在此基础上,总结出两种数据流组织方案,用于确定各种硬件和性能约束下硬件架构的最佳配置。它提供了多级跟踪结果,可作为 RTL 设计的黄金模型。此外,基于广义硬件模型的 SimuNN 可以评估各种量化、数据流组织和硬件配置下的硬件性能。在此基础上,总结出两种数据流组织方案,用于在各种硬件和性能约束下确定硬件架构的最佳配置。
更新日期:2020-06-01
down
wechat
bug