BindMark - XmlBeans marshalling code revisited
XmlBeans 2.0.0 marshalling code has been improved by user's suggestion. Using the correct function in base class, the marshalling figures (both time and memory) have improved dramatically by up to 99%...
BindMark - Javolution 3.6.4 evaluated
Javolution 3.6.4 has been evaluated. The main points:
First time unmarshalling has footprint larger by 6-11%.
Level unmarshalling of medium sized XML is 13% faster.
Level unmarshalling of large s...
BindMark - Javolution 3.5.3 evaluated
New version of Javolution (release 3.5.3) has been evaluated. This release features improvements in both time and memory:
Marshalling - the first time has improved by 33-63 percents, while consecut...
BindMark - JSX (first time) evaluation
After being in no-support state for a few months, the JSX library is back on track (at the same URL and with the same versioning policy). The latest version is 2.2.5.0 and comes with 30-day trial lice...
BindMark - corrections and evaluation of C24 3.2 and JaxMe 0.5
A major update to the BindMark test schema. It was pointed by Dennis Sosnoski that xs:integer is represented by most schema-based libraries as BigInteger. The implementation of BigInteger.toString() i...
BindMark - JiBX 1.0RC1 evaluated
The new JiBX 1.0RC1 release has been evaluated. Alongside the binaries, this evaluation also included removing hand-written classes and binding definition and moved to generating classes and binding d...
BindMark - JiBX toolkit evaluated
The toolkit of JiBX 1.0RC1 has been evaluated.
The two important tools for JiBX are binding generator and xsd2jibx. The binding generator takes a set of existing classes and creates a default bindi...
BindMark - Skaringa r3p7 evaluated
The latest Skaringa r3p7 release was evaluated. As this was largely bug fix release, the performance was not affected, as well as the scoring for all data sets....
BindMark - Betwixt correct use of shared Registry
The Betwixt marshalling and unmarshalling code has been corrected to make use of shared Registry for the reflection information. This allows memory improvements of 45KB per marshalling / unmarshalling...
BindMark - Castor 0.9.7 evaluated
The latest Castor 0.9.7 version has been evaluated. This version provides improvements for unmarshalling time and footprint for larger data sets (-30% in time and -18% in memory footprint), along with...
BindMark - JAXB 2.0 RI August snapshot evaluated
The last time JAXB 2.0 RI was evaluated in April, and since then it showed significant improvements in both time and memory.
The marshalling has been improved by 56% on small sets and by 86-91% (...
BindMark - Betwixt first time evaluation
A new library, Betwixt, has been evaluated.
Both time and memory performance of Betwixt are among the worst of all libraries, especially for unmarshaling (takes four times as much memory as the secon...
BindMark - JaxMe first time evaluation
A new library, JaxMe, was evaluated for the first time. JaxMe is an open-source implementation (Apache) of JAXB 1.0 and is completely inter-operable with JAXB 1.0 (using generated classes, marshaling ...
BindMark - Javolution 3.4.0 evaluated
New version of Javolution has been evaluated. Both marshaling and unmarshaling have been improved in time performance on after-the-first (level) runs. The marshaling is 14-21% faster (across all datas...
BindMark - XMLBeans new release and XmlIO first evaluation
XMLBeans 2.0.0 official release has been evaluated. A few fluctuation from the beta (back in February), but overall the performance remains the same. The marshaller is still dominated (90% in both tim...