db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Daniel John Debrunner (JIRA)" <j...@apache.org>
Subject [jira] Commented: (DERBY-2328) Reduce monitor contention in SinglePool
Date Tue, 13 Feb 2007 15:13:05 GMT

    [ https://issues.apache.org/jira/browse/DERBY-2328?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12472702

Daniel John Debrunner commented on DERBY-2328:

I  think you are proposing that the LockFactory has a createCompatibilitySpace method that
returns some interface, say CompatibilitySpace. Then all methods in the lock api that take
'Object compatibilitySpace' would change to 'CompatibilitySpace compatibilitySpace'. Without
the specific interface the api could be become subject for bugs if a object not created by
createCompatibilitySpace is passed in.
With this approach I don't see any need to specifiy/require anything about equals for CompatibilitySpace,
but maybe I'm missing something.

I think this is a good approach, the existing code valued footprint over multi-cpu performance
(and maybe the existing approach did not save footprint due to the memory used by the hashtable).
Seems this approach only needs one new class, the interface CompatibilitySpace, and then the
existing LockSpace would implement it.

> Reduce monitor contention in SinglePool
> ---------------------------------------
>                 Key: DERBY-2328
>                 URL: https://issues.apache.org/jira/browse/DERBY-2328
>             Project: Derby
>          Issue Type: Sub-task
>          Components: Performance, Services
>    Affects Versions:
>            Reporter: Knut Anders Hatlen
>            Priority: Minor
> When multiple threads enter the lock manager, there might be contention on SinglePool's
monitor. The contention should be reduced in order to improve scalability.

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

View raw message