[jira] [Commented] (MAHOUT-606) Parallelize non-distributed ALSWRFactorizer

["Georgi Stanev (Commented) (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/MAHOUT-606?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13165233#comment-13165233 ] 

Georgi Stanev commented on MAHOUT-606:
--------------------------------------

Got a little problem with that patch.

Got an assertion in the solver at solve on line 158 that both vector lengths are unequal. What did helped is instead building the featureVectors (line 156) in a while with iterator, build it in foreach (like line 194) using the PreferenceArray (line 149). Should be the same but there is obviously a problem with the iterator.

Just want to share those thoughts for anyone who have problems with this.
                
> Parallelize non-distributed ALSWRFactorizer
> -------------------------------------------
>
>                 Key: MAHOUT-606
>                 URL: https://issues.apache.org/jira/browse/MAHOUT-606
>             Project: Mahout
>          Issue Type: Improvement
>          Components: Collaborative Filtering
>    Affects Versions: 0.5
>            Reporter: Sebastian Schelter
>            Assignee: Sean Owen
>             Fix For: 0.5
>
>         Attachments: MAHOUT-606.patch
>
>
> Add multithreading to ALSWRFactorizer so that all available cores are used for the computation of the factorization

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira