You are viewing a plain text version of this content. The canonical link for it is here.
Posted to ojb-dev@db.apache.org by Armin Waibel <ar...@code-au-lait.de> on 2003/02/05 02:14:33 UTC
bad karma: OJB performance slow down
Hi all,
since the last version of OJB (0.9.8) I trace the
performance of the CVS version.
Here my results running the 'perfomance2' target
(10 threads handle 2000 objects per thread insert/fetch/delete)
[sec] [sec]
date PB-api ODMG-api
------------------------------
0.9.8 21,2 53,5
01/20 17,7 42,5
01/25 17,8 40,4
01/31 17,6 41,1
02/04 21,4 42,4
The refactoring made after the 0.9.8 release
boost the PB-api performance about 17%, the
ODMG-api about 21%. After changes made
by Thomas (I suppose) the ODMG performance was
enhanced about 25% compared with the 0.9.8
version. This are great results.
But some of the changes made in the PB-api between
01/31 and today seems to be performance killer!
Decrease the performance of the PB-api on the 0.9.8 level.
Does anybody has an idea how we could find out
the 'bad' lines of code slows down the performance?
regards,
Armin
Re: bad karma: OJB performance slow down
Posted by Sanne de Roever <sa...@newfoundland.nl>.
This profiling tool is free:
http://www.hp.com/products1/unix/java/hpjmeter/index.html
But I don't think it's LGPL.
----- Original Message -----
From: "Armin Waibel" <ar...@code-au-lait.de>
To: "OJB Developers List" <oj...@db.apache.org>
Sent: Wednesday, February 05, 2003 12:09 PM
Subject: Re: bad karma: OJB performance slow down
> Hi Thomas,
>
> ----- Original Message -----
> From: "Thomas Mahler" <th...@web.de>
> To: "OJB Developers List" <oj...@db.apache.org>
> Sent: Wednesday, February 05, 2003 7:12 AM
> Subject: Re: bad karma: OJB performance slow down
>
>
> > Hi Armin,
> > It's a very good idea to track the performance impacts of all changes
> we
> > are doing!
> > We could add the performance target to the junit target. If we dod so
> > this test will be included in the nightly gump builds.
>
> +1
> indeed this will help to make OJB a high quality project ;-)
>
> Currently I'm writing a small (very small) multithreaded
> performance test framework based on Performance2.java.
> This makes it easy to implement multithreaded performance
> tests for our API's (PB,ODMG,JDO).
> Further on it will be possible to compare OJB with other
> persistence layer - more info soon.
>
> regards,
> Armin
>
>
> >
> >
> > cheers,
> > Thomas
> >
> > Armin Waibel wrote:
> > > Hi all,
> > >
> > > since the last version of OJB (0.9.8) I trace the
> > > performance of the CVS version.
> > > Here my results running the 'perfomance2' target
> > > (10 threads handle 2000 objects per thread insert/fetch/delete)
> > >
> > > [sec] [sec]
> > > date PB-api ODMG-api
> > > ------------------------------
> > > 0.9.8 21,2 53,5
> > > 01/20 17,7 42,5
> > > 01/25 17,8 40,4
> > > 01/31 17,6 41,1
> > > 02/04 21,4 42,4
> > >
> > > The refactoring made after the 0.9.8 release
> > > boost the PB-api performance about 17%, the
> > > ODMG-api about 21%. After changes made
> > > by Thomas (I suppose) the ODMG performance was
> > > enhanced about 25% compared with the 0.9.8
> > > version. This are great results.
> > >
> > > But some of the changes made in the PB-api between
> > > 01/31 and today seems to be performance killer!
> > > Decrease the performance of the PB-api on the 0.9.8 level.
> > >
> > > Does anybody has an idea how we could find out
> > > the 'bad' lines of code slows down the performance?
> > >
> > > regards,
> > > Armin
> > >
> > >
> >
> > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: ojb-dev-unsubscribe@db.apache.org
> > > For additional commands, e-mail: ojb-dev-help@db.apache.org
> > >
> > >
> >
> >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: ojb-dev-unsubscribe@db.apache.org
> > For additional commands, e-mail: ojb-dev-help@db.apache.org
> >
> >
> >
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: ojb-dev-unsubscribe@db.apache.org
> For additional commands, e-mail: ojb-dev-help@db.apache.org
>
Re: bad karma: OJB performance slow down
Posted by Armin Waibel <ar...@code-au-lait.de>.
Hi Thomas,
----- Original Message -----
From: "Thomas Mahler" <th...@web.de>
To: "OJB Developers List" <oj...@db.apache.org>
Sent: Wednesday, February 05, 2003 7:12 AM
Subject: Re: bad karma: OJB performance slow down
> Hi Armin,
> It's a very good idea to track the performance impacts of all changes
we
> are doing!
> We could add the performance target to the junit target. If we dod so
> this test will be included in the nightly gump builds.
+1
indeed this will help to make OJB a high quality project ;-)
Currently I'm writing a small (very small) multithreaded
performance test framework based on Performance2.java.
This makes it easy to implement multithreaded performance
tests for our API's (PB,ODMG,JDO).
Further on it will be possible to compare OJB with other
persistence layer - more info soon.
regards,
Armin
>
>
> cheers,
> Thomas
>
> Armin Waibel wrote:
> > Hi all,
> >
> > since the last version of OJB (0.9.8) I trace the
> > performance of the CVS version.
> > Here my results running the 'perfomance2' target
> > (10 threads handle 2000 objects per thread insert/fetch/delete)
> >
> > [sec] [sec]
> > date PB-api ODMG-api
> > ------------------------------
> > 0.9.8 21,2 53,5
> > 01/20 17,7 42,5
> > 01/25 17,8 40,4
> > 01/31 17,6 41,1
> > 02/04 21,4 42,4
> >
> > The refactoring made after the 0.9.8 release
> > boost the PB-api performance about 17%, the
> > ODMG-api about 21%. After changes made
> > by Thomas (I suppose) the ODMG performance was
> > enhanced about 25% compared with the 0.9.8
> > version. This are great results.
> >
> > But some of the changes made in the PB-api between
> > 01/31 and today seems to be performance killer!
> > Decrease the performance of the PB-api on the 0.9.8 level.
> >
> > Does anybody has an idea how we could find out
> > the 'bad' lines of code slows down the performance?
> >
> > regards,
> > Armin
> >
> >
>
> ---------------------------------------------------------------------
> > To unsubscribe, e-mail: ojb-dev-unsubscribe@db.apache.org
> > For additional commands, e-mail: ojb-dev-help@db.apache.org
> >
> >
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: ojb-dev-unsubscribe@db.apache.org
> For additional commands, e-mail: ojb-dev-help@db.apache.org
>
>
>
Re: bad karma: OJB performance slow down
Posted by Thomas Mahler <th...@web.de>.
Hi Armin,
It's a very good idea to track the performance impacts of all changes we
are doing!
We could add the performance target to the junit target. If we dod so
this test will be included in the nightly gump builds.
cheers,
Thomas
Armin Waibel wrote:
> Hi all,
>
> since the last version of OJB (0.9.8) I trace the
> performance of the CVS version.
> Here my results running the 'perfomance2' target
> (10 threads handle 2000 objects per thread insert/fetch/delete)
>
> [sec] [sec]
> date PB-api ODMG-api
> ------------------------------
> 0.9.8 21,2 53,5
> 01/20 17,7 42,5
> 01/25 17,8 40,4
> 01/31 17,6 41,1
> 02/04 21,4 42,4
>
> The refactoring made after the 0.9.8 release
> boost the PB-api performance about 17%, the
> ODMG-api about 21%. After changes made
> by Thomas (I suppose) the ODMG performance was
> enhanced about 25% compared with the 0.9.8
> version. This are great results.
>
> But some of the changes made in the PB-api between
> 01/31 and today seems to be performance killer!
> Decrease the performance of the PB-api on the 0.9.8 level.
>
> Does anybody has an idea how we could find out
> the 'bad' lines of code slows down the performance?
>
> regards,
> Armin
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: ojb-dev-unsubscribe@db.apache.org
> For additional commands, e-mail: ojb-dev-help@db.apache.org
>
>