I think the reason of comparing two programs in two different languages is exactly to have an idea of how expressive one language can be to generate a performing program given its inputs and outputs are the same, the way on which the code is written is the second factor being evaluated. The benchmark game is the perfect exemple of this, anyone can write a better version of a program and improve the language "score" until no one will be able to improve it. At this moment we can see how the language design impacts the resulting program. This should be the only way to compare languages.