DUET: Boosting Deep Neural Network Efficiency on Dual-Module Architecture
DUET: Boosting Deep Neural Network Efficiency on Dual-Module Architecture
复制标题
DOI：
10.1109/micro50266.2020.00066
复制
发表时间：
2020-10
期刊：
影响因子：
--
通讯作者：
Liu Liu-Liu;Zheng Qu;Lei Deng;Fengbin Tu;Shuangchen Li;Xing Hu;Zhenyu Gu;Yufei Ding;Yuan Xie
中科院分区：
文献类型：
--
作者：
Liu Liu-Liu;Zheng Qu;Lei Deng;Fengbin Tu;Shuangchen Li;Xing Hu;Zhenyu Gu;Yufei Ding;Yuan Xie
深度神经网络（DNN）一直在推动机器学习应用的主流。然而，由于各种DNN模型的计算密集型和内存密集型执行模式，在具有严格延迟要求和能量约束的现代硬件上部署DNN具有挑战性。我们提出了一种算法架构协同设计，以提高DNN的执行效率。利用DNN中非线性激活函数的抗噪声能力，我们提出了双模块处理，该处理使用从原始DNN层学习的近似模块来计算不敏感的激活。因此，我们可以节省昂贵的计算和不必要的敏感激活的数据访问。然后，我们设计了一个Executor-Speculator双模块架构，支持平衡执行和内存访问减少。在可接受的模型推理质量下降的情况下，我们的加速器设计可以为计算限制的卷积神经网络（CNN）和内存限制的递归神经网络（RNN）实现2.24倍的加速和1.97倍的能效提升。
Deep Neural Networks (DNNs) have been driving the mainstream of Machine Learning applications. However, deploying DNNs on modern hardware with stringent latency requirements and energy constraints is challenging because of the compute-intensive and memory-intensive execution patterns of various DNN models. We propose an algorithm-architecture co-design to boost DNN execution efficiency. Leveraging the noise resilience of nonlinear activation functions in DNNs, we propose dual-module processing that uses approximate modules learned from original DNN layers to compute insensitive activations. Therefore, we can save expensive computations and data accesses of unnecessary sensitive activations. We then design an Executor-Speculator dual-module architecture with support for balance execution and memory access reduction. With acceptable model inference quality degradation, our accelerator design can achieve 2.24x speedup and 1.97x energy efficiency improvement for compute-bound Convolutional Neural Networks (CNNs) and memory-bound Recurrent Neural Networks (RNNs).