Hi-Speed DNN Training with Espresso: Unleashing the Full Potential of Gradient Compression with Near-Optimal Usage Strategies

Zhuang Wang, Haibin Lin, Yibo Zhu, T. S. Eugene Ng

Abstract: Gradient compression (GC) is a promising approach to addressing the communication bottleneck in distributed deep learning (DDL). It saves the communication time, but also incurs additional computation overheads. The training throughput of compression-enabled DDL is determined by the compression strategy, including whether to compress each tensor, the type of compute resources (e.g., CPUs or GPUs) for compression, the communication schemes for compressed tensor, and so on. However, it is challenging to find the optimal compression strategy for applying GC to DDL because of the intricate interactions among tensors. To fully unleash the benefits of GC, two questions must be addressed: 1) How to express any compression strategies and the corresponding interactions among tensors of any DDL training job? 2) How to quickly select a near-optimal compression strategy?In this paper, we propose Espresso to answer these questions. It first designs a decision tree abstraction to express any compression strategies and develops empirical models to timeline tensor computation, communication, and compression to enable Espresso to derive the intricate interactions among tensors. It then designs a compression decision algorithm that analyzes tensor interactions to eliminate and prioritize strategies and optimally offloads compression from GPUs to CPUs. Experimental evaluations show that Espresso can improve the training throughput over the start-of-the-art compression-enabled system by up to 77% for representative DDL training jobs. Moreover, the computational time needed to select the compression strategy is measured in milliseconds, and the selected strategy is only a few percent from optimal.