Mailing-List: contact commits-help@cassandra.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@cassandra.apache.org
Date: Thu, 1 May 2014 15:09:18 +0000 (UTC)
From: "Jonathan Ellis (JIRA)" <jira@apache.org>
To: commits@cassandra.apache.org
Message-ID: <JIRA.12711657.1398914426204.213652.1398956958410@arcas>
In-Reply-To: <JIRA.12711657.1398914426204@arcas>
References: <JIRA.12711657.1398914426204@arcas>
Subject: [jira] [Commented] (CASSANDRA-7128) Upgrade NBHM
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/CASSANDRA-7128?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13986655#comment-13986655 ] 

Jonathan Ellis commented on CASSANDRA-7128:
-------------------------------------------

I thought cleaning out the links and attachments would be more work than a new issue, but go for it

> Upgrade NBHM
> ------------
>
>                 Key: CASSANDRA-7128
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-7128
>             Project: Cassandra
>          Issue Type: Improvement
>            Reporter: T Jake Luciani
>            Assignee: T Jake Luciani
>            Priority: Trivial
>              Labels: performance
>             Fix For: 2.1 beta2
>
>         Attachments: 7128.txt, Screen Shot 2014-04-30 at 11.07.22 PM.png
>
>
> AbstractRowResolver uses a NBHM for each read request.  
> Profiler flagged this as a bottleneck since the init() call creates a AtomicReferenceFieldUpdater which is stored in a synchronized collection.
> A NBHS is most certainly overkill for such a short lived object.  And turns out switching it to a CHS in my tests yields a ~5-10% read boost.


--
This message was sent by Atlassian JIRA
(v6.2#6252)