Why are you changing all of those pumps and rads? You are making things too hard to manage. Control the water temperature to remove it as a variable. Use a single powerful pump and use valves to adjust flow rates to preset points. Then monitor CPU performance. Not rocket science to devise; very difficult to do well.
Testing 3x to avoid errors in mounting? That would presume that you control mounting pressure in the first place (a good idea eh?).
The real rub is that the thermal paste will kick you in the nut sac once you get the resolution to see its effects.
If I can actually get my PC-50 finished I will try and get my testbed set up. No promises though
