It need more testing on both sides.
If clock stays at maximum during benchmark, how can the score be lower?
Also leaving everything at stock and setting AC and DC load line to something around 0.80, 0.60 ohms or lower, gives a nice undervolt leaving cpu voltage at auto. At least this didn't trigger any of the bugs from undervolting on my board.
Just keep clear CMOS at hand when trying AC and DC ohms values.