DL2: A Deep Learning-Driven Scheduler for Deep Learning Clusters
This article proposes and implements a DL-driven scheduler for DL clusters, targeting global training job expedition by dynamically resizing resources allocated to jobs, and implements DL techniques to enable dynamic resource scaling in DL jobs on MXNet.
Abstract
Efficient resource scheduling is essential for maximal utilization of expensive deep learning (DL) clusters. Existing cluster schedulers either are agnostic to machine learning (ML) workload characteristics, or use scheduling heuristics based on operators’ understanding of particular ML framework and workload, which are less efficient or not general enough. In this article, we show that DL techniques can be adopted to design a generic and efficient scheduler. Specifically, we propose DL<inline-formula><tex-math notation="LaTeX">$^2$</tex-math><alternatives><mml:math><mml:msup><mml:mrow/><mml:mn>2</mml:mn></mml:msup></mml:math><inline-graphic xlink:href="peng-ieq2-3052895.gif"/></alternatives></inline-formula>, a DL-driven scheduler for DL clusters, targeting global training job expedition by dynamically resizing resources allocated to jobs. DL<inline-formula><tex-math notation="LaTeX">$^2$</tex-math><alternatives><mml:math><mml:msup><mml:mrow/><mml:mn>2</mml:mn></mml:msup></mml:math><inline-graphic xlink:href="peng-ieq3-3052895.gif"/></alternatives></inline-formula> advocates a joint supervised learning and reinforcement learning approach: a neural network is warmed up via offline supervised learning based on job traces produced by the existing cluster scheduler; then the neural network is plugged into the live DL cluster, fine-tuned by reinforcement learning carried out throughout the training progress of the DL jobs, and used for deciding job resource allocation in an online fashion. We implement DL<inline-formula><tex-math notation="LaTeX">$^2$</tex-math><alternatives><mml:math><mml:msup><mml:mrow/><mml:mn>2</mml:mn></mml:msup></mml:math><inline-graphic xlink:href="peng-ieq4-3052895.gif"/></alternatives></inline-formula> on Kubernetes and enable dynamic resource scaling in DL jobs on MXNet. Extensive evaluation shows that DL<inline-formula><tex-math notation="LaTeX">$^2$</tex-math><alternatives><mml:math><mml:msup><mml:mrow/><mml:mn>2</mml:mn></mml:msup></mml:math><inline-graphic xlink:href="peng-ieq5-3052895.gif"/></alternatives></inline-formula> outperforms fairness scheduler (i.e., DRF) by 44.1 percent and expert heuristic scheduler (i.e., Optimus) by 17.5 percent in terms of average job completion time.