Deep Learning Assisted Resource Partitioning for Improving Performance on Commodity Servers

Ruobing Chen, Jinping Wu, Haosen Shi, Yusen Li, Haiyan Yin, Shanjiang Tang, Xiaoguang Liu, Gang Wang

Abstract: In this paper, we introduce a deep reinforcement learning (DRL) framework for solving the problem of partitioning LLC and memory bandwidth coordinately in an end-to-end manner. To this end, we formulate the problem as a markov decision process and utilize DRL algorithm to derive the optimal partition. To avoid the extensive cost of training the policy on physical server, we present a model-based solution, where a reward prediction model is leveraged to train the partitioning policy offline. To construct a precise reward prediction model, we introduce a novel representation for the partitioning scheme, where graph convolutional networks (GCN) is employed to represent the LLC partition as a bipartite graph so that those heterogeneous but identical partitions could result in the same representations and thus eases the prediction task. Experimental results show that our framework is able to make immediate and very competitive partitioning decisions, which improves the system performance with significant margins compared to the baseline without resource partitioning and the state-of-the-art single resource partitioning solutions.