I received the response from Mark Pollack on the Spring.Net forum which gives a pretty definitive answer (Mark is Co Lead on the Spring.Net project):
"Hi,
There is a lock around retrieving/creating singleton objects and non singleton objects, prototypes, are effecdtively 'newed' within the calling thread. I've revived some tests around this area, ConcurrentObjectFactoryTests, which had a config file for ages but no test impl using it, as well as in DefaultListableObjectFactoryPerfTests.
There were some microbench marks that I did a while ago when there was attention on this in the Java world between Spring and Guice. (The cutely named "Semi Useless Benchmark").
http://www.javalobby.org/articles/gu...ource=archives
http://www.jroller.com/Solomon/entry...e_improvements
The ".NET ported" classes from the benchmark are inside DefaultListableObjectFactoryPerfTests. I don't have any performance data for Spring.NET. The community has reported issues with concurrency in the past (quite a while ago) which were fixed and I'm unaware of any specific issues in this area.
HTH,
Mark"