BML: A High-performance, Low-cost Gradient Synchronization Algorithm for DML Training

Wang, Songtao, Li, Dan, Cheng, Yang, Geng, Jinkun, Wang, Yanshu, Wang, Shuai, Xia, Shu-Tao, Wu, Jianping

Neural Information Processing Systems 

In distributed machine learning (DML), the network performance between machines significantly impacts the speed of iterative training. In this paper we propose BML, a new gradient synchronization algorithm with higher network performance and lower network cost than the current practice. BML runs on BCube network, instead of using the traditional Fat-Tree topology. BML algorithm is designed in such a way that, compared to the parameter server (PS) algorithm on a Fat-Tree network connecting the same number of server machines, BML achieves theoretically 1/k of the gradient synchronization time, with k/5 of switches (the typical number of k is 2∼4). Experiments of LeNet-5 and VGG-19 benchmarks on a testbed with 9 dual-GPU servers show that, BML reduces the job completion time of DML training by up to 56.4%.