Energy-Efficient Mapping for a Network of DNN Models at the Edge
TimeTuesday, December 7th6:00pm - 7:00pm PST Event Type
DescriptionThis paper describes a novel framework for executing a network of DNNs on commercial-off-the-shelf devices. The scenario consists of two devices connected by a wireless network: an energy and performance-limited user-end device (U), and an energy-unconstrained cloudlet (C). The goal is to distribute the computation of DNNs between U and C to minimize the energy consumption of U while considering the variability in wireless delay and the overhead of executing models in parallel. Experiments demonstrate significant improvements in energy consumption of U where U is an NVIDIA Jetson Nano and C is a Dell workstation with Titan Xp GPU.