[ 
https://issues.apache.org/jira/browse/HAMA-704?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Thomas Jungblut updated HAMA-704:
---------------------------------

    Attachment: HAMA-704_1337.patch

First scratch in HAMA-704_1337.patch, everything within maven and locally works 
so far (even pagerank!).
I will profile this tomorrow or a bit later today to see if it has given any 
major improvement on the RAM usage.

Still I'm not completely sure about many things, do you think we should open a 
review board thread for it? Please let me know about suggestions and honest 
opinions you have.

Also a small drawback: we can't skip reading inactive vertices every iteration 
because the values file must be rewritten using all vertex values (otherwise 
there are obviously strange results). However, I added the possibility to do so 
by a strategy pattern if needed or we have a better idea how to not rewrite the 
whole values every iteration.
Also for aggregators to work correctly, I had to replay messages between the 
supersteps (between the master and the slave synchronization). Maybe we can get 
a better synchronization strategy that doesn't clear the queues so we don't 
have to loop-back them again for the next step.
                
> 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: Critical
>             Fix For: 0.6.1
>
>         Attachments: HAMA-704_1337.patch, HAMA-704.patch-v1, 
> hama-704_v05.patch, HAMA-704-v2.patch, localdisk.patch, mytest.patch, 
> patch.txt, patch.txt, removeMsgMap.patch
>
>
> <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

Reply via email to