[rules-users] Add/remove objects from working memory are very CPU intensive

[Einat Idan]

Hi Mark,

Thanks, these is good news!
My product is planned to be GA in 4-5 months. When do you recommend shifting
from 3.0.1 (that's the version I currently use) to 3.1? When do you expect
3.1 to be a stable version?

Best regards,
Einat

On 4/12/07, Mark Proctor <mproctor at codehaus.org> wrote:
>
>  I'm just finalising the last bit, so any day now. with any luck over the
> weekend or monday.
>
> Mark
> Einat Idan wrote:
>
> Michael,
> Thanks for your reply.
>
> When is the next milestone expected?
>
> On 4/12/07, Michael Neale <michael.neale at gmail.com > wrote:
> >
> > definately upgrade to latest 3.0.x version (3.0.6).
> >
> > Also, those methods are were most of the work happens, its a common
> > misconception that all the work happens lazily when you call "fire all
> > rules" but that is not the case, as you assert each object, it propagates
> > through the RETE network, so that is normal to see the time spent there for
> > lots of data.
> >
> > you can also try the trunk version if you like, its certainly got some
> > improvements, but the next milestone (if you can wait) will be more worth
> > your time.
> >
> > Michael.
> >
> >  On 4/12/07, Einat Idan <idan.einat at gmail.com> wrote:
> >
> > > Hi,
> > >
> > > I encountered a serious performance problem using Jboss Rules 3.0.1.
> > > The process was executed on a pretty strong machine - a DL350 4 cpu RedHat
> > > machine. The process was using about 100% CPU and I used a profiler to see
> > > what's going on:
> > >
> > > My application adds/removes objects to/from the working memory of a
> > > stateful rule session quite intensively (2000-3000 per sec), though the
> > > intensive actions were related to a single rule session and only a few extra
> > > rule sessions existed simultaneously. It turned out that about 7-10% of CPU
> > > was consumed per a single add/remove operation. More specifically,
> > > ReteooWorkingMemory.doRetract() and ReteooWorkingMemory.doAssertObject()
> > > were the major consumers. I would expect a basic operation like this to be
> > > significantly less CPU intensive.
> > >
> > > Would you please provide more information, is my benchmark too
> > > ambitious? Do you recommend an upgrade to version 3.0.6? 3.1? If so,
> > > please elaborate what were the performance improvements.
> > >
> > > Best regards,
> > > Einat Idan
> > >
> > >
> > >
> > >
> > >
> > >  _______________________________________________
> > > rules-users mailing list
> > > rules-users at lists.jboss.org
> > > https://lists.jboss.org/mailman/listinfo/rules-users
> > >
> > >
> >
> > _______________________________________________
> > rules-users mailing list
> > rules-users at lists.jboss.org
> > https://lists.jboss.org/mailman/listinfo/rules-users
> >
> >
> ------------------------------
>
> _______________________________________________
> rules-users mailing list
> rules-users at lists.jboss.orghttps://lists.jboss.org/mailman/listinfo/rules-users
>
>
>
> _______________________________________________
> rules-users mailing list
> rules-users at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/rules-users
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.jboss.org/pipermail/rules-users/attachments/20070412/2dcacdc0/attachment.html