When looking at benchmark numbers of any kind, always be wary of what a benchmark truly is: simulated relative tasks and not real world tasks. While there is value in a benchmark as it is a way to somewhat measure performance between different components, but keep in mind the tests the benchmark program runs may have nothing to do with what you do in real use of the component.

Personally, I’ve always believed the philosophy, that “benchmark = lie”. I’m not going to go so far as to say this is true, rather emphasize real world performance over test scores.

Think about this: suppose there is a program “Test My Graphics Card” which is the defacto standard for measuring graphics card performance. Don’t you think graphics card manufacturers would optimize their cards to score well on this program, regardless of real world performance? I’m not trying to be cynical here, rather proposing a realistic scenario.

All I am saying here, is to simply think twice before letting a benchmark score determine the worth or effectiveness of hardware or software. By no means is a benchmark the “holy grail” of performance measurement.