hama-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Edward J. Yoon (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HAMA-745) ListVerticesInfo doesn't work
Date Fri, 17 May 2013 11:47:18 GMT

    [ https://issues.apache.org/jira/browse/HAMA-745?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13660591#comment-13660591

Edward J. Yoon commented on HAMA-745:

My idea for reducing memory usage, was use of sorted vertices and sequential (or serial) processing[1]:

Edward J. Yoon added a comment - 30/Jan/13 07:25
I'm thinking about external merge sort http://en.wikipedia.org/wiki/External_sorting
Edward J. Yoon added a comment - 04/Feb/13 06:28
Once HAMA-723 is done, we can sort messages by vertex ID. Then, we can read messages of vertex
and perform vertex.compute() sequentially.

And, Thomas Jungblut implemented serial processing and DiskVerticesInfo but, the problem was
that PartitioningRunner creates unsorted partitions.

Do you work on this issue?

1. https://issues.apache.org/jira/browse/HAMA-704?focusedCommentId=13566257&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-13566257

> ListVerticesInfo doesn't work
> -----------------------------
>                 Key: HAMA-745
>                 URL: https://issues.apache.org/jira/browse/HAMA-745
>             Project: Hama
>          Issue Type: Bug
>            Reporter: Edward J. Yoon
>            Assignee: Edward J. Yoon
>             Fix For: 0.6.1
>         Attachments: HAMA-745.patch, HAMA-745_v02.patch
> Just tried to use ListVerticesInfo instead of DiskVerticesInfo but doesn't work correctly.
The reason is that addVertex() changes the same object. It need to reinitializing vertex object.

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

View raw message