Reliable Allreduce and Broadcast Interface for distributed machine learning
Geek Repo:Geek Repo
Github PK Tool:Github PK Tool
MaJun-cn opened this issue 5 years ago · comments
please,I want to know more details about using the rabit to implement program communication in a distributed cluster.
please, I want to know how to use it.
@MaJun-cn Can you elaborate what do communication protocol you are looking for? rabit implemented MPI all-reduce and broadcast interface with plain socket in C++. Tracker by default using python to keep track of all participants.