Here's the inherent problem with comparing temperatures while testing various configurations..
I ran these tests a day apart with identical configurations. The only major variable is the ambient temperature in the room, which shows a 4°C difference which yields, not surprisingly, a 4°C difference in the results.
We can pull the covers off, flip the fans, change fan profiles, but if we aren't listing our ambient temperatures accurately, comparing data with each other becomes a lie of omission.
If anyone posts any pseudo-scientific results regarding a configuration change that they believe to be better, please also post accurate ambient temperatures. Check a few sources; the thermostat if you're in an apartment, a thermometer left on a non-conducting surface such as a cloth, a case sensor left dangling outside the case, etc.
This is absolutely critical. If you don't include an ambient temperature and you post that your GPUs are running at 50°C and all you did was change the fan profile, someone is going to see that information on this forum, attempt to repeat the experiment and fail.
Just a friendly reminder to include as many critical details as possible.