The longevity is a good point, but if those numbers are derived from users running PassMark on those servers, I would suspect that they don't keep running the s

The longevity is a good point, but if those numbers are derived from users running PassMark on those servers, I would suspect that they don't keep running the software on their server they purchased some years ago: From my understanding, people would be more likely to run this software on a recently-purchased computer to evaluate its performance.