Hi Jimmy,
Nice improvements with buffer reuse. I have found that this is a very
effective way to improve performance.
Given that VTD-XML does not perform namespace processing until actually
traversing it may be more meaningful in these test cases to actually
switch it off, if possible, for the other APIs.
Another, possibly simple, approach (beyond a real example test case) may
be to compare VTD against DOM with traversal of the whole infoset.
I am surprised that the deferred DOM only improved memory usage between
1 to 2 of the original document size. IMHO this is a little disappointing.
Paul.
Jimmy Zhang wrote:
> I tried to benchmark everything again with and without deferred node
> expansion
> and the new benchmark numbers are posted on the vtd-xml web site
> http://www.ximpleware.com/benchmark1.html
>
> basically it is to compare the performance of the following
>
> VTD-XML 1.5 (to be released soon)
> VTD-XML 1.5 with buffer reuse turned on
> Xerces DOM 2.7.1 (deferred node expansion)
> Xerces DOM 2.7.1 (fully expanded)
> Xerces SAX 2.7.1
> Piccolo
> XPP Pull.
>
>
>
>
> ----- Original Message ----- From: "Paul Sandoz" <Paul.Sandoz_at_Sun.COM>
> To: <users_at_fi.dev.java.net>
> Sent: Thursday, January 26, 2006 1:39 AM
> Subject: Xerces DOM deferred node expansion <was> Re: VTD speed
>
>
>> Jimmy Zhang wrote:
>>
>>> I was trying to rerun the benchmark with deferred node
>>> expansion set to false, is there reference on hwo to do this??
>>> I use setFeature on DOMBuilderFactory but has no success...
>>>
>>> Any suggestions?
>>
>>
>> Try this:
>>
>> DocumentBuilderFactory bf = ...
>> bf.setAttribute("http://apache.org/xml/features/dom/defer-node-expansion",
>>
>> false).
>>
>> Paul.
>>
>> --
>> | ? + ? = To question
>> ----------------\
>> Paul Sandoz
>> x38109
>> +33-4-76188109
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe_at_fi.dev.java.net
>> For additional commands, e-mail: users-help_at_fi.dev.java.net
>>
>>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe_at_fi.dev.java.net
> For additional commands, e-mail: users-help_at_fi.dev.java.net
>
--
| ? + ? = To question
----------------\
Paul Sandoz
x38109
+33-4-76188109