You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hama.apache.org by "Bhathiya Supun (JIRA)" <ji...@apache.org> on 2014/09/01 03:07:21 UTC
[jira] [Commented] (HAMA-891) Implement greedy heuristic
partitioning algorithm
[ https://issues.apache.org/jira/browse/HAMA-891?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14116957#comment-14116957 ]
Bhathiya Supun commented on HAMA-891:
-------------------------------------
As metioned mailing list we would like to work on this issue. [~udanax] , Can you assign this issue to me?
> Implement greedy heuristic partitioning algorithm
> -------------------------------------------------
>
> Key: HAMA-891
> URL: https://issues.apache.org/jira/browse/HAMA-891
> Project: Hama
> Issue Type: New Feature
> Components: graph
> Reporter: Edward J. Yoon
> Labels: gsoc2014
>
> See http://www.charlesneedham.com/pubs/166151/kdd325-stanton.pdf
> {quote}Conclusion:
> We have demonstrated that simple, one-pass streaming graph partitioning heuristics can dramatically improve the edge-cut in distributed graphs. Our best performing heuris- tic is the linear weighted variant of the greedy algorithm. This is a simple and effective preprocessing step for large graph computation systems, as the data must be loaded onto the cluster any way.
> ...{quote}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)