[jboss-jira] [JBoss JIRA] Commented: (JBRULES-2240) True modify
Geoffrey De Smet (JIRA)
jira-events at lists.jboss.org
Sun Aug 9 05:01:29 EDT 2009
[ https://jira.jboss.org/jira/browse/JBRULES-2240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12479646#action_12479646 ]
Geoffrey De Smet commented on JBRULES-2240:
-------------------------------------------
This can potentially give a huge performance boost, especially in statefull environments.
One such case this the 2 accumulates in a single rule case:
<conan> when the first accumulate is updated
<conan> the second accumulate has it's memory wiped
<ge0ffrey> ouch
<conan> because the first one says I'm re-calcuating this
<conan> so go to any child nodes for all my propagations and remove them
<ge0ffrey> So drools doesn't detect that the second accumulate does not depend on the first? That there's no correlation?
<ge0ffrey> That the result of the first accumulate is not used in calculation for the second?
<conan> doesn't matter
<conan> it's a chain of tuples remember
<conan> if you break the parent
<conan> then everything below is retracted
<ge0ffrey> would true modify fix this?
<conan> think about it a second
<conan> it's a chain of tuples
<conan> the second, cannot exist without hte first
<conan> if I need to re-calculate the first
<conan> the nthe second cannot exist.
<conan> ge0ffrey: true modify will improve this, yes.
<ge0ffrey> because then the second would not need to be retracted if the first is modified
<conan> in true modify if the first one was true and contineus to be true
<conan> whiel it's evaluated
<conan> it doesn't destroy the second one
<ge0ffrey> unless its correlated
<conan> if hte first one is false
<conan> the second one will need to be destroyed
But there are many more cases where true modify can have huge impact on performance.
> True modify
> -----------
>
> Key: JBRULES-2240
> URL: https://jira.jboss.org/jira/browse/JBRULES-2240
> Project: Drools
> Issue Type: Feature Request
> Security Level: Public(Everyone can see)
> Components: drools-core (expert)
> Affects Versions: 5.0.1.FINAL
> Reporter: Geoffrey De Smet
> Assignee: Mark Proctor
> Fix For: 5.1.0.M1
>
>
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
More information about the jboss-jira
mailing list