While not a performance test report, I can make the following observation: Java retains the performance of Java, and Lua retrains the performance of Lua.
So, the only thing that is new with JNLua is the overhead of transitions between Lua and Java. This includes activities such as JNI transitions, setting up a pcall, dynamic method dispatch, argument checking, etc.
JNLua does perform optimizations where reasonable and possible, e.g. by caching dynamic method dispatch results for future calls. However, most of the transition overhead is technically unavoidable given JNLua’s goal of providing an access model to native Lua where bad arguments or out of memory conditions do not cause a JVM crash.
So, if you are happy with the Java performance and the Lua performance of your code (each taken for itself), the questions to answer really is how much transition overhead between the two is added into the mix.
While not a direct answer to your question, I still hope you find this answer useful.