[vanderZwan]

> so it is almost certainly getting the wrong answers

Not checking the correctness of the output sounds like a pretty bad oversight for a benchmark