Adaptive Workload Distribution for Accuracy-aware DNN Inference on Collaborative Edge Platforms
This work addresses performance and accuracy challenges for DNN inference on collaborative edge platforms, representing an incremental improvement over existing workload distribution methods.
The paper tackled the problem of optimizing DNN inference on heterogeneous edge devices by proposing an adaptive workload distribution method that jointly considers device heterogeneity and accuracy-performance trade-offs, achieving an average gain of 41.52% in performance and 5.2% in accuracy compared to state-of-the-art strategies.
DNN inference can be accelerated by distributing the workload among a cluster of collaborative edge nodes. Heterogeneity among edge devices and accuracy-performance trade-offs of DNN models present a complex exploration space while catering to the inference performance requirements. In this work, we propose adaptive workload distribution for DNN inference, jointly considering node-level heterogeneity of edge devices, and application-specific accuracy and performance requirements. Our proposed approach combinatorially optimizes heterogeneity-aware workload partitioning and dynamic accuracy configuration of DNN models to ensure performance and accuracy guarantees. We tested our approach on an edge cluster of Odroid XU4, Raspberry Pi4, and Jetson Nano boards and achieved an average gain of 41.52% in performance and 5.2% in output accuracy as compared to state-of-the-art workload distribution strategies.