[ 
https://issues.apache.org/jira/browse/LUCENE-2041?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12786932#action_12786932
 ] 

Israel Tsadok commented on LUCENE-2041:
---------------------------------------

This has turned into a complete refactoring of the class, but I'd like to 
comment that the original intent of this patch was a bit misguided: docFreq() 
is not called during normal use of ParallelMultiSearcher. What does get called 
is docFreqs() on the sub-searchers. This is done in 
MultiSearcher.createWeight(), and is done sequentially.

I created issue LUCENE-2128 to suggest a solution to this.

> Complete parallelizaton of ParallelMultiSearcher
> ------------------------------------------------
>
>                 Key: LUCENE-2041
>                 URL: https://issues.apache.org/jira/browse/LUCENE-2041
>             Project: Lucene - Java
>          Issue Type: Improvement
>          Components: Search
>    Affects Versions: 2.9
>            Reporter: Joey Surls
>            Assignee: Uwe Schindler
>            Priority: Minor
>             Fix For: 3.0
>
>         Attachments: LUCENE-2041-final.patch, LUCENE-2041-refactor.patch, 
> LUCENE-2041-refactor.patch, LUCENE-2041.patch, LUCENE-2041.patch, 
> LUCENE-2041.patch, LUCENE-2041.patch, LUCENE-2041.patch, LUCENE_2041.patch
>
>   Original Estimate: 1h
>  Remaining Estimate: 1h
>
> ParallelMultiSearcher is parallel only for the method signatures of 'search'.
> Part of a query process calls the method docFreq(). There was a TODO comment 
> to parallelize this. Parallelizing this method actually increases the 
> performance of a query on multiple indexes, especially remotely.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: java-dev-h...@lucene.apache.org

Reply via email to