This is a very simple, rough, and probably bad test of RNG wrapping-unwrapping performance. Hotspotting makes it challenging, because the implemented pattern is far from typical.
On my MBP with:
java version "1.8.0_66"
Java(TM) SE Runtime Environment (build 1.8.0_66-b17)
Java HotSpot(TM) 64-Bit Server VM (build 25.66-b17, mixed mode)
and:
2.3 GHz Intel Core i7
The program output is:
> run
[info] Running RNGWrappingPerformance
scala.util.Random 760ms 1.3157894736842105E7/s
java.util.Random 708ms 1.4124293785310734E7/s
MersenneTwister 555ms 1.801801801801802E7/s
RandomAdaptor MT 597ms 1.6750418760469012E7/s
----
scala.util.Random => java.util.Random 744ms 1.3440860215053763E7/s
java.util.Random => scala.util.Random 739ms 1.3531799729364008E7/s
scala.util.Random => RandomGenerator 624ms 1.6025641025641026E7/s