Dorylus: Affordable, Scalable, and Accurate GNN Training with Distributed CPU Servers and Serverless Threads.
John ThorpeYifan QiaoJonathan EyolfsonShen TengGuanzhou HuZhihao JiaJinliang WeiKeval VoraRavi NetravaliMiryung KimGuoqing Harry XuPublished in: OSDI (2021)
Keyphrases
- scalable distributed
- distributed systems
- training set
- single point of failure
- high quality
- cooperative
- high scalability
- computationally efficient
- distributed environment
- commodity hardware
- multi threaded
- load balance
- parallel computing
- training process
- fault tolerant
- high accuracy
- databases
- computing environments
- training examples
- fault tolerance
- test set
- distributed processing
- lightweight
- online learning
- wide area network
- general purpose
- multi agent
- remote server
- neural network