Also check this one: http://dartogreniyorum.blogspot.ch/2013/05/performance-optim...
and later with Simd:
http://dartogreniyorum.blogspot.ch/2013/05/dart-beats-java-i...
Btw, I think even with the modifications made to Java version mentioned in the link you gave, I think latest version of dart now beats Java.
Edit: Author of blog post also compared Richards and Tracer benchmarks: http://bonovox.be/blog/dart-vs-java-richards-and-tracer/
However it's worth noting that deltablue has been used to benchmark polymorphism in Smalltalk, Self, the JVM, and V8.
Some of Nickolay's numbers are from a version of the code written at Sun labs, presumably for the very purpose of benchmarking the JVM.
I think it's safe to conclude that in many cases the DartVM already performs similarly to the JVM, albeit with a much better start up time.
Charles Nutter has some interesting follow up comments here:
http://blog.headius.com/2013/05/on-languages-vms-optimizatio...
Basically concluding that it's not static typing that matters, but predictable types.
IMO DeltaBlue is a terrible benchmark typology anyway. Benchmarks have to be localized for them to mean anything; and even then, they don't mean much 90% of the time. Dart's pretty neat (editing a book about it atm), but this kind of language zealotry is just silly. Dart is hardly a silver bullet. And many of the truly interesting features (mixins, isolates, fancier mirrors) are yet to be implemented or incomplete. Definitely keep an eye out for Dart, but don't fall all over yourself thinking that it outperforms Java (whatever that even means).