RankMap: Priority-Aware Multi-DNN Manager for Heterogeneous Embedded Devices

TL;DR

RankMap is a novel priority-aware management system for multi-DNN workloads on heterogeneous embedded devices, significantly improving throughput and prioritization while preventing DNN starvation.

Contribution

It introduces RankMap, a new system that efficiently manages multiple DNNs by combining stochastic exploration and performance estimation tailored for heterogeneous embedded systems.

Findings

Achieves 3.6x higher throughput than existing methods.

Prevents DNN starvation under heavy workloads.

Improves prioritization of specific DNNs by 57.5x.

Abstract

Modern edge data centers simultaneously handle multiple Deep Neural Networks (DNNs), leading to significant challenges in workload management. Thus, current management systems must leverage the architectural heterogeneity of new embedded systems to efficiently handle multi-DNN workloads. This paper introduces RankMap, a priority-aware manager specifically designed for multi-DNN tasks on heterogeneous embedded devices. RankMap addresses the extensive solution space of multi-DNN mapping through stochastic space exploration combined with a performance estimator. Experimental results show that RankMap achieves x3.6 higher average throughput compared to existing methods, while preventing DNN starvation under heavy workloads and improving the prioritization of specified DNNs by x57.5.