How does the Dozer Java API work?
In general, display performance is the same either using XML configuration or Java API.
Performance will depend on your application (use cases), in general Orika performs better than Dozer or other mapping (reflection based frameworks) because it uses behind the scenes bytecode generation.
IMO, when you have to render HTML and access remote data storage, the performance of the map can be messy.
Here are some steps http://bit.ly/pJ7n6t
source to share
I just stumbled upon GeDA - a really fast bean mapping framework according to http://inspire-software.com/confluence/display/GeDA/Benchmarks and with Spring integration support.
Had 2 classes with 4 simple fields - String, String, Date, boolean. 1,000,000 conversions of such an object took 100ms with GeDA and 1000ms with Orika.
Knowing that Orika is 10x faster than Dozer, you can imagine how Geada vs. Dozer performs
source to share