JIT -LUA emprovements

[publicocean0]

Based on stackoverflow link (https://stackoverflow.com/questions/19001241/big-execution-time-difference-between-java-lambda-vs-anonymous-class/46767495#46767495) i m writing this one for helping to emproving JIT.
Syncerelly i m upset for this result. Functional interface is just a remapping from paramters to parameters based on upper clousure context(if the root functional context is null ,as in the most cases, it is equivalent to a inline replacement), instead anonymous class in addition use the class loader for create a new class... so i expect functional is faster because is pratically a inline replacement in the most cases.All the lambda theory is not used if you dont use mathematics for deducting this info based on theory. It seams jit is not making that analysis for understanding it(pay attention ...this analisis could be present at static bytecode generation time before jit is executing, in fact there are info present at static code formulation that jit can use for fastest jit compilation).I suspect computer science must emprove a lot in future.
In addition for JIT emprovements there are also all a series of compilation infos calculated at bytecode generation are characterised by be unique and not modifiable by class definition. The JIT perfomance is not only base on the code generated by also in what time is is produced the best code because also it has a impact in the sotware execution perfomance.The behavior of JIT and the form of bytecode must not be based on JIT developers opinions or based "so before was" criterion but based on mathematic criterion producing the lowest "energy" for producing the best JIT result.