[
https://issues.apache.org/jira/browse/SOLR-5201?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13757536#comment-13757536
]
Tommaso Teofili commented on SOLR-5201:
---------------------------------------
here's a draft patch:
https://github.com/tteofili/lucene-solr/compare/apache:trunk...solr-5201.patch
_AnalysisEngines_ are initialized inside _UIMAUpdateRequestProcessorFactories_
together with a _JCasPool_ to better handle multiple concurrent requests.
My benchmarks (ran 'ant clean test -Dtests.multiplier=100' with and without the
above patch) show execution of
_UIMAUpdateRequestProcessorTest#testMultiplierProcessing_ is ~10 times faster
and less memory consumptive (~240MB saved over ~650MB heap)
> UIMAUpdateRequestProcessor should reuse the AnalysisEngine
> ----------------------------------------------------------
>
> Key: SOLR-5201
> URL: https://issues.apache.org/jira/browse/SOLR-5201
> Project: Solr
> Issue Type: Improvement
> Components: contrib - UIMA
> Affects Versions: 4.4
> Reporter: Tommaso Teofili
> Assignee: Tommaso Teofili
> Fix For: 4.5, 5.0
>
> Attachments: SOLR-5201-ae-cache-every-request_branch_4x.patch,
> SOLR-5201-ae-cache-only-single-request_branch_4x.patch
>
>
> As reported in http://markmail.org/thread/2psiyl4ukaejl4fx
> UIMAUpdateRequestProcessor instantiates an AnalysisEngine for each request
> which is bad for performance therefore it'd be nice if such AEs could be
> reused whenever that's possible.
--
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
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
