A Sum-of-Ratios Multi-Dimensional-Knapsack Decomposition for DNN Resource Scheduling

TL;DR

This paper introduces a novel resource scheduling framework for DNN training that leverages the layered structure of DNN jobs, using a sum-of-ratios multi-dimensional knapsack approach to improve job completion times in large-scale clusters.

Contribution

It presents a new analytical model for DNN resource scheduling, and develops an efficient algorithm based on sum-of-ratios decomposition with strong performance guarantees.

Findings

Significantly reduces DNN job completion times

Outperforms existing scheduling methods in experiments

Provides a scalable and effective scheduling solution

Abstract

In recent years, to sustain the resource-intensive computational needs for training deep neural networks (DNNs), it is widely accepted that exploiting the parallelism in large-scale computing clusters is critical for the efficient deployments of DNN training jobs. However, existing resource schedulers for traditional computing clusters are not well suited for DNN training, which results in unsatisfactory job completion time performance. The limitations of these resource scheduling schemes motivate us to propose a new computing cluster resource scheduling framework that is able to leverage the special layered structure of DNN jobs and significantly improve their job completion times. Our contributions in this paper are three-fold: i) We develop a new resource scheduling analytical model by considering DNN's layered structure, which enables us to analytically formulate the resource scheduling optimization problem for DNN training in computing clusters; ii) Based on the proposed performance analytical model, we then develop an efficient resource scheduling algorithm based on the widely adopted parameter-server architecture using a sum-of-ratios multi-dimensional-knapsack decomposition (SMD) method to offer strong performance guarantee; iii) We conduct extensive numerical experiments to demonstrate the effectiveness of the proposed schedule algorithm and its superior performance over the state of the art.