EdgeWise: Energy-efficient CNN Computation on Edge Devices under Stochastic Communication Delays
EdgeWise: Energy-efficient CNN Computation on Edge Devices under Stochastic Communication Delays
复制标题
DOI：
10.1145/3530908
复制
发表时间：
2022-04
期刊：
影响因子：
--
通讯作者：
Mehdi Ghasemi;Daler N. Rakhmatov;Carole-Jean Wu;S. Vrudhula
中科院分区：
文献类型：
--
作者：
Mehdi Ghasemi;Daler N. Rakhmatov;Carole-Jean Wu;S. Vrudhula
本文提出了一个框架，使卷积神经网络（cnn）在边缘设备上的节能执行。该框架由一对通过无线网络连接的边缘设备组成：一个性能和能量受限的设备D作为数据的第一个接收者，一个能量不受限的设备N作为D的加速器。设备D在考虑网络延迟固有的不确定性和数据传输所涉及的开销的同时，动态地决定如何分配工作负载，以最大限度地减少其能耗。通过采用马尔可夫决策过程的数据驱动建模框架来解决这些挑战，D在O(1)时间内咨询最优策略以做出逐层分配决策。作为一种特殊情况，本文还提出了一种线性时间动态规划算法，在假设网络延迟在整个应用程序执行过程中是恒定的情况下，立即找到最优的层分配。所提出的框架在一个由树莓派3作为D和NVIDIA Jetson TX2作为n组成的平台上进行了演示。与完全在D和n上执行cnn的替代方案相比，能耗平均提高了31%和23%。
This article presents a framework to enable the energy-efficient execution of convolutional neural networks (CNNs) on edge devices. The framework consists of a pair of edge devices connected via a wireless network: a performance and energy-constrained device D as the first recipient of data and an energy-unconstrained device N as an accelerator for D. Device D decides on-the-fly how to distribute the workload with the objective of minimizing its energy consumption while accounting for the inherent uncertainty in network delay and the overheads involved in data transfer. These challenges are tackled by adopting the data-driven modeling framework of Markov Decision Processes, whereby an optimal policy is consulted by D in O(1) time to make layer-by-layer assignment decisions. As a special case, a linear-time dynamic programming algorithm is also presented for finding optimal layer assignment at once, under the assumption that the network delay is constant throughout the execution of the application. The proposed framework is demonstrated on a platform comprised of a Raspberry PI 3 as D and an NVIDIA Jetson TX2 as N. An average improvement of 31% and 23% in energy consumption is achieved compared to the alternatives of executing the CNNs entirely on D and N. Two state-of-the-art methods were also implemented and compared with the proposed methods.