Kirill,
The cost of object allocation is taken into account during unmarshalling, since this is where the object hierarchy is created. I don't quite understand why this would be counted again during marshalling. If you're going to count the creation of the object hierarchy for the marshalling benchmark, just make sure your consistent across all of the products tested. For example, add the cost of org.jvnet.bindmark.benchmark.test.Jaxb2Test.createHierarchy (365 kB, 10,812 alloc.) durinig marshalling for JAXB 2.0.
James
Kirill Grouchnikov <kirillcool_at_yahoo.com> wrote:
Giac,
I still consider the initialization of the binding library
as part of one-time marshalling. Although it happens in the
creation phase, it affects the overall footprint of the
process. Compared to other libraries, you not only "pay the
price" of object allocation of your own hierarchy (which is
ignored in the tests), but also the allocation of XmlBeans.
That's why i don't ignore it in the results.
Kirill
--- JAXB User wrote:
> Kirill,
>
> The XMLBeans functions you mentioned are not part of a
> one-time marshalling cost, they are part of the creation
> of the object hierarchy, which as you explained before,
> is not counted in the marshalling benchmark. I believe
> all the one-time marshalling cost is captured under the
> org.apache.xmlbeans.XmlObject.save call, since the first
> run takes quite a bit more memory than each subsequent
> iteration.
>
> It would be interesting to measure creation performance
> as a separate benchmark, but I'm sure there's no shortage
> of other stuff to keep you busy! :-)
>
> Regards,
>
> Giac
>
> Kirill Grouchnikov wrote:
> Giac,
>
> Thanks for the suggestions. I'll test the XMLBeans on it.
> However, it seems that you have missed the one-time cost
> of
> initializing the XMLBeans. In
> org.jvnet.bindmark.benchmark.marshal.MarshalerThread.run
> go
> to
>
org.jvnet.bindmark.benchmark.test.XmlBeansTest.createHierarchy
> and to
>
org.jvnet.bindmark.gen.xmlbeans.OrganizationDocument$Factory.newInstance.
>
> You'll see the constructor of XmlBeans object which takes
> additional 3.190KB and 25.570 objects.
>
> I'll update the reports over the weekend
> Kirill
>
> --- JAXB User wrote:
>
> > Kirill,
> >
> > The numbers I get using your method outlined below are
> > actually still quite a bit different. So far I've
> > benchmarked Sun's JAXB 2.0 reference implementation
> (from
> > 9/19/05) and XMLBeans (v2), several times with the same
> > results. One important change I made in the XMLBeans
> > test is how marshalling is performed. Instead of
> > creating a string out of the XML being marshalled and
> > then serializing it, I use the save method on their
> base
> > object like this:
> >
> > ((XmlObject)hierarchy).save(baos);
> >
> > So the lower numbers I get (i.e. 6595kb versus 101238kb
> > and 140ms versus 25864ms) would be explained by this.
> > Would it be possible for you to make this change on
> your
> > end?
> >
> > To get the Level time number, I multiple the average
> time
> > reported to System.out by the number of iterations
> (i.e.
> > 1000 iterations for small, 100 for medium, and 10 for
> > large data sets).
> >
> > My setup is as follows:
> >
> > JDK 1.5.0_03
> > Windows XP SP 2
> > Dell Lititude D600 with 1.86Ghz Intel Pentium M
> > processor and 2GB RAM
> > JProfiler 4.0.9
> > Below are the results I get. Regards,
> >
> > GIac
> >
> >
> > Marshaling
> >
> >
> >
> > Time (ms)
> >
> > Memory (kb)
> >
> > Memory (objects)
> >
> > Startup
> >
> > Level
> >
> > startup
> >
> > level
> >
> > startup
> >
> > level
> >
> > JAXB 2.0
> >
> > Small data sets
> >
> > 6.2
> >
> > 122
> >
> > 7579
> >
> > 5349
> >
> > 89689
> >
> > 53041
> >
> > Medium data sets
> >
> > 21
> >
> > 179
> >
> > 7740
> >
> > 15976
> >
> > 91132
> >
> > 133759
> >
> > Large data sets
> >
> > 65
> >
> > 140
> >
> > 8239
> >
> > 6595
> >
> > 96118
> >
> > 63474
> >
> > XMLBeans
> >
> > Small data sets
> >
> > 39
> >
> > 223
> >
> > 378
> >
> > 31307
> >
> > 1920
> >
> > 51116
> >
> > Medium data sets
> >
> > 144
> >
> > 273
> >
> > 629
> >
> > 11530
> >
> > 8561
> >
> > 13540
> >
> > Large data sets
> >
> > 190
> >
> > 130
> >
> > 1501
> >
> > 3972
> >
> > 35008
> >
> > 34290
> >
> >
> >
> > Unmarshaling
> >
> >
> >
> > Time (ms)
> >
> > Memory (kb)
>
=== message truncated ===
__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe_at_jaxb.dev.java.net
For additional commands, e-mail: users-help_at_jaxb.dev.java.net
---------------------------------
Yahoo! for Good
Click here to donate to the Hurricane Katrina relief effort.