You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@mxnet.apache.org by ha...@apache.org on 2018/01/23 19:18:19 UTC

[incubator-mxnet] branch master updated: Docs changes for launch.py usage (#8000)

This is an automated email from the ASF dual-hosted git repository.

haibin pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git


The following commit(s) were added to refs/heads/master by this push:
     new a840ae5  Docs changes for launch.py usage (#8000)
a840ae5 is described below

commit a840ae58ef0f4fa659aa044e448ff23033664e8c
Author: Anirudh Subramanian <an...@gmail.com>
AuthorDate: Tue Jan 23 11:18:15 2018 -0800

    Docs changes for launch.py usage (#8000)
    
    * Add additional information about launch.py
    
    * Add multi devices documentation for launch.py
    
    * Add more info about args
    
    * Update multi_devices.md
    
    * Add -s explanation
    
    * Correct launch comment
    
    * Correction
    
    * Add docs
---
 docs/faq/multi_devices.md | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/docs/faq/multi_devices.md b/docs/faq/multi_devices.md
index c79d1f8..9bd582c 100644
--- a/docs/faq/multi_devices.md
+++ b/docs/faq/multi_devices.md
@@ -148,6 +148,14 @@ Note that here we
 - `-n` number of worker nodes to run on
 - `-H` the host file which is required by `ssh` and `mpi`
 - `--kv-store` use either `dist_sync` or `dist_async`
+- `-s` number of server nodes to run on
+- If the `-s` argument is not passed, it will keep the number of servers same as number of workers
+- The launch.py script tries to cycle through the hosts file to launch the servers and workers. For example,
+  let's say you have `5` hosts in the hosts file and you passed n as `3`(and nothing for s).
+  The script will launch a total of `3` server processes, one each for the first three hosts and
+  launch a total of `3` worker processes, one each for the fourth, fifth and first host.
+- If the hosts file has exactly `n` number of worker nodes which is passed as an argument with `-n`, it will launch
+  a server process and a worker process on each of the `n` hosts.
 
 
 ### Synchronize Directory

-- 
To stop receiving notification emails like this one, please contact
haibin@apache.org.