NeuPart: Using Analytical Models to Drive Energy-Efficient Partitioning of CNN Computations on Cloud-Connected Mobile Clients
NeuPart: Using Analytical Models to Drive Energy-Efficient Partitioning of CNN Computations on Cloud-Connected Mobile Clients
复制标题
DOI：
10.1109/tvlsi.2020.2995135
复制
发表时间：
2019-05
影响因子：
2.8
通讯作者：
Susmita Dey Manasi;F. S. Snigdha;S. Sapatnekar
中科院分区：
文献类型：
--
作者：
Susmita Dey Manasi;F. S. Snigdha;S. Sapatnekar
卷积神经网络（CNN）上的数据处理给能量受限的移动的平台带来了沉重的负担。本文通过在客户端上的原位处理和云中的卸载计算之间划分CNN计算来优化移动的客户端上的能量。制定了一个新的分析CNN能量模型，捕获了基于ASIC的深度学习加速器的原位计算的所有主要组件。该模型的基准对测得的硅数据。分析框架用于确定运行时客户端和云之间的最佳能量划分点。在标准CNN拓扑上，分区计算被证明可以在完全基于云的计算或完全原位计算上为客户端提供显着的节能。例如，在80 Mbps有效比特率和0.78 W传输功率下，AlexNet [SqueezeNet]的最佳分区在完全基于云的计算中节省了高达52.4% [73.4%]的能量，在完全原位计算中节省了27.3% [28.8%]的能量。
Data processing on convolutional neural networks (CNNs) places a heavy burden on energy-constrained mobile platforms. This article optimizes energy on a mobile client by partitioning CNN computations between in situ processing on the client and offloaded computations in the cloud. A new analytical CNN energy model is formulated, capturing all major components of the in situ computation, for ASIC-based deep learning accelerators. The model is benchmarked against measured silicon data. The analytical framework is used to determine the optimal energy partition point between the client and the cloud at runtime. On standard CNN topologies, partitioned computation is demonstrated to provide significant energy savings on the client over a fully cloud-based computation or fully in situ computation. For example, at 80 Mbps effective bit rate and 0.78 W transmission power, the optimal partition for AlexNet [SqueezeNet] saves up to 52.4% [73.4%] energy over a fully cloud-based computation and 27.3% [28.8%] energy over a fully in situ computation.