[
https://issues.apache.org/jira/browse/HAMA-704?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13548363#comment-13548363
]
Edward J. Yoon commented on HAMA-704:
-------------------------------------
The parseMessages() method parses 'GraphJobMessage' object to extract vertexID
and the messages sent from other vertices in the previous iteration, and
finally returns the HashMap <VertexID, Messages>. Then,
{code}
for(Map.entry<VertexID, MessagesList> e : msgMap) {
vertices.get(e.getKey()).compute(e.getValue());
}
{code}
Thus, the Previous messages and new messages to be sent to next iteration
should be all in memory, during computing every vertices.
My simple idea is the use of local disk. Then, memory usage will be reduced by
between 40% to 50%.
> Optimization of memory usage during message processing
> ------------------------------------------------------
>
> Key: HAMA-704
> URL: https://issues.apache.org/jira/browse/HAMA-704
> Project: Hama
> Issue Type: Improvement
> Components: graph
> Reporter: Edward J. Yoon
> Assignee: Edward J. Yoon
> Priority: Minor
> Fix For: 0.6.1
>
>
> <vertex, message> map seems consume a lot of memory. We should figure out an
> efficient way to reduce memory.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
