Thanks, guys.
I have another mainboard waterblock test coming up. I'll use identical thermal paste for each set of blocks there.

There's a semi-unfair bias towards the Bitspower block in the test above. But it's only semi-unfair because the fact that the BP blocks come with AS5 is a real advantage. It's just problematic because I have two variables in one test...
Anyway, I noticed this mistake too late and don't want to hold up other reviews by retesting everything.