[weld-issues] [JBoss JIRA] Commented: (WELD-785) "Minor refactor and fix test to be more robust" commit to weld/core/master causes a big performance decline

[Pete Muir (JIRA)]

    [ https://issues.jboss.org/browse/WELD-785?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12569438#comment-12569438 ] 

Pete Muir commented on WELD-785:
--------------------------------

http://hudson.qa.jboss.com/hudson/view/Weld/job/perf-weld-numberguess-stress-single/144/artifact/report/log.txt contains the results, which look like it is back up to the correct point.

Martin, reopen if you see any problems with that run :-)

> "Minor refactor and fix test to be more robust" commit to weld/core/master causes a big performance decline
> -----------------------------------------------------------------------------------------------------------
>
>                 Key: WELD-785
>                 URL: https://issues.jboss.org/browse/WELD-785
>             Project: Weld
>          Issue Type: Bug
>          Components: Conversations, Scopes & Contexts
>            Reporter: Martin Gencur
>            Assignee: Pete Muir
>            Priority: Critical
>             Fix For: 1.1.0.CR1
>
>
> This commit causes a big performance decline when the application is being accessed from higher number of sessions. Now the stress test can only reach 6500 sessions at most. Removing this commit gets the max number of sessions up to 12600. This is mainly caused by heap memory not being freed sufficiently.

-- 
This message is automatically generated by JIRA.
-
For more information on JIRA, see: http://www.atlassian.com/software/jira