[
https://issues.apache.org/jira/browse/OPENJPA-1872?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12928288#action_12928288
]
Rick Curtis commented on OPENJPA-1872:
--------------------------------------
Committed revision 1031110 to trunk.
> Optimize CacheMap read performance
> ----------------------------------
>
> Key: OPENJPA-1872
> URL: https://issues.apache.org/jira/browse/OPENJPA-1872
> Project: OpenJPA
> Issue Type: Improvement
> Components: performance
> Affects Versions: 2.1.0
> Reporter: Rick Curtis
> Assignee: Rick Curtis
> Priority: Minor
> Fix For: 2.1.0
>
>
> While digging through a profile I noticed that for every CacheMap.get(..)
> call where there is a hit, I was seeing two underlying ConcurrentHashMap.get
> method calls. This is because we always look to see if a given key exists in
> the pinned map before looking in the main map. As the code stands today, it
> is optimized for the existence of entries in the pinned map and I think that
> is an incorrect assumption.
> With this JIRA I'm going to optimize the get(...), containsKey(...), and
> containsValue(...) to check the main map for results before looking at the
> pinned map. In a micro benchmark I observed approximately a 20% improvement
> in the time required to call get(...) when there is a hit.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
