[ 
https://issues.apache.org/jira/browse/SAMZA-245?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14072304#comment-14072304
 ] 

Chris Riccomini commented on SAMZA-245:
---------------------------------------

Also, did some CPU sampling to see how CPU cycles are spent after this patch. 
About 40% of CPU is spent inside SystemStreamPartition.hashCode inside 
OffsetManager.update. I think we can optimize this in a separate patch, but 
it's good to know that there's still room to grow.

By commenting out the OffsetManager.update line, I was able to do 10,000,000 
messages in 9 seconds on my little Macbook Air, which puts us over 1 million 
messages/sec. So, this looks like the next thing to target.

> Improve SystemConsumers performance
> -----------------------------------
>
>                 Key: SAMZA-245
>                 URL: https://issues.apache.org/jira/browse/SAMZA-245
>             Project: Samza
>          Issue Type: Bug
>          Components: container
>    Affects Versions: 0.6.0
>            Reporter: Chris Riccomini
>            Assignee: Chris Riccomini
>             Fix For: 0.8.0
>
>         Attachments: SAMZA-245-1.patch, SAMZA-245-3.patch, SAMZA-245-4.patch, 
> SAMZA-245-5.patch, SAMZA-245-6.patch, SAMZA-245-7.patch, SAMZA-245.0.patch, 
> org.apache.samza.test.performance.TestSamzaContainerPerformance.SAMZA-245-3.html,
>  org.apache.samza.test.performance.TestSamzaContainerPerformance.master.html
>
>
> As part of SAMZA-220, a more radical patch was proposed. This patch appears 
> to improve SystemConsumers' performance pretty significantly, while also 
> reducing its complexity. The decision was made to move this change into the 
> 0.8.0 release, rather than the 0.7.0 release, since it's a fairly risky 
> change.
> This ticket is to explore updating SystemConsumers to eliminate almost all 
> loops in order to increase performance in the Samza container.



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

Reply via email to