Benchmarks are meaningless unless properly taken.
One drive can operate twice as fast as another of the exact same model, and both can still be within manufacturing specs.One benchmark chart can be completely different from another, but both can be completely accurate. My point is not th consider such benchmarks at such a fine granularity or to go through them with a fine tooth comb; statistics are meaningless when the method to gather that data is flawed.
Also, <who> "certifies" the drives for RAID?