[TuG3] Algorithms for Photonic Computing

3:00 PM - 3:15 PM JST(6:00 AM - 6:15 AM UTC)

[TuG3-5] Accelerating Distributed Machine Learning with AllReduce Reconfiguration Based on Optical Circuit Switching

Zilong Ye^1,2, ○Philip N. Ji¹, Ting Wang¹ (1. NEC Laboratories America (United States of America), 2. California State University Los Angeles (United States of America))

Keywords:

Photonics for neuromorphic computing and deep learning

We propose to apply optical circuit switching to enable dynamic AllReduce reconfiguration for accelerating distributed machine learning. With simulated annealing-based optimization, the proposed AllReduce reconfiguration approach achieves 31% less average training time than existing solutions.