Just to give a correct interpretation of the SMART Data and the benchmark and clear up the misunderstandings revealed in some answers, comments and the question:
The SMART Data shows a normal degradation of an approx. three years old hard disk.
The benchmark shows the normal behavior of a spinning hard drive.
Usually the linear density of magnetic areas holding a bit is almost constant on the whole platter. Therefore the outermost track contains more bits than the innermost track. With one rotation of the platter the read/write head can read/write more bits on the outermost track than the innermost track in the same time.
Example:
Rotations: 1/second
Bit-density: 100 Bits/inch
Radius of the outermost track: 2 inches -> track length: 12,6 inches
Radius of the innermost track: 1 inch -> track length: 6,3 inches
Bits on the outermost track: 1260 B
Bits on the innermost track: 630 B
Read/write rate outermost track: 1260 B/s
Read/write rate innermost track: 630 B/s
As a result the read rate (blue) and the write rate (pink) slightly decreases from the outermost tracks on the left (≈ 84 MB/s) to the innermost tracks on the right (≈ 45 MB/s).
The average access time (the time it takes before the drive can actually transfer data from a random track.) is also normal (≈ 17 ms). We see only 4 outliers (marked with red circles) in 1000 samples. One sample has an access time of 105 ms and 3 samples of ≈ 90 ms. All other samples show an access time between 7 ms and 30 ms.
The thin grey lines between the small green dots (every dot is one of the 1000 samples) should show the chronological succession of the access time benchmarking test. I emphasized 6 consecutive accesses with a thick green line.
