I didn't bother to do a super accurate average, because let's be honest, being a few percent off for this makes no difference to my end point. Hit my 1800X with the CPU burner function of Furmark are let it stabilize for a few minutes before moving on to the next pump speed.

The measurements are the difference between ambient temperature and the CPU core temperature. The case and AIO fan speed was fixed to prevent that from being a factor.

~42.1 C delta at 10% pump PWM

~40.4 C delta at 20% pump PWM

~39.3 C delta at 30% pump PWM

~39.0 C delta at 40% pump PWM

~38.8 C delta at 50% pump PWM

~38.5 C delta at 60% pump PWM

~38.3 C delta at 70% pump PWM

~38.0 C delta at 80% pump PWM (max speed)

So in case it isn't abundantly clear, there is virtually no benefit to running the pump at high RPM, or even much above 30% PWM. Running it at super low PWM is also not a great idea, there is virtually no audible difference between 10% PWM and 30% PWM.

All of the complaints about the new pump being louder at max RPM, for me, is a little silly. There is very little difference in performance between low and high RPMs on the pump. In my opinion, the significantly higher noise of 50% PWM isn't even close to worth the slight improvement to core temperature.

On the flip side, running the pump too slow, around 20% or less, has a pretty significant decrease in performance. To my judgement, this is also best avoided.

In my opinion, the sweet spot of my unit is 30% PWM. Anything faster is pretty clearly hitting diminishing returns. However, the pump is next to inaudible even at 40% PWM, so objectively there isn't a huge difference between running either unless you cared about pump wear, at least with my set up.

The moral of the story, even if the pump CAN get louder at max pump speed, doesn't mean you should be running it anywhere near it, and you'd be best off running it next to silent anyways.