ignite-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Ilya Lantukh (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (IGNITE-2579) Investigate HashMap.Node[] allocations from GridCacheMvccManager$3
Date Fri, 19 Feb 2016 12:08:18 GMT

    [ https://issues.apache.org/jira/browse/IGNITE-2579?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15154148#comment-15154148
] 

Ilya Lantukh commented on IGNITE-2579:
--------------------------------------

In our benchmarks HashSet size is always equal to 1. I presume this is also true for the most
typical use-cases.
Adjusted constructor parameters accordingly.

> Investigate HashMap.Node[] allocations from GridCacheMvccManager$3
> ------------------------------------------------------------------
>
>                 Key: IGNITE-2579
>                 URL: https://issues.apache.org/jira/browse/IGNITE-2579
>             Project: Ignite
>          Issue Type: Sub-task
>          Components: cache
>    Affects Versions: 1.5.0.final
>            Reporter: Vladimir Ozerov
>            Assignee: Ilya Lantukh
>             Fix For: 1.6
>
>
> *Problem* 
> See GridCacheMvccManager.addFuture() method. We create a weird HashSet there with internal
table size == 5. Can we have something more efficient here?
> *Proposed solution*
> Need to run single get-put benchmarks and check usual size of this collection. If it
is often equal to 1, then instead of allocating the whole collection, we'd better to have
a singleton first and expand to collection if there are more elements. 
> Please pay attention that collection usually used as monitor in some synchronized blocks.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message