Efficient All-Reduce for Distributed DNN Training in Optical Interconnect Systems.

Fei Dai Yawen Chen Zhiyi Huang Haibo Zhang Fangfang Zhang

Published in: PPoPP (2023)

Keyphrases

distributed systems
open systems
building blocks
cooperative
training process
fault tolerant
lightweight
intelligent systems
peer to peer
high speed
management system
data sets
training set
expert systems
data intensive
heterogeneous systems
distributed network
distributed object