I don't have a Mac Pro but I have a 10-core Vega 64 iMac Pro and a M1 Max MBP 16. Although the iMac Pro was often described as a quiet machine if I push it hard with computationally-intensive effects in DaVinci Resolve (e.g, temporal/spatial noise reduction and tracked masks), it is pretty loud. The iMac Pro is quiet *relative* to my 2017 iMac 27, and the fans spin up more slowly, but the iMac Pro still can get pretty loud. I also have a 2019 top-spec MacBook Pro 16 which can also be pretty loud under similar conditions.
By contrast the M1 Max 16 is very quiet. In iStat Menus the fans indicate 0 rpm until it is pushed really hard for several minutes, then they go to about 1500 rpm. At that point they are difficult to hear unless my ear is about 3 inches from the keyboard. As the compute-intensive task continues for several more minutes the rpm increases to about 1900 rpm, at which point I can hear it in a normal position if everything in the room is quiet. After a few more min it hits 2500 rpm, where it's easier to hear but still much quieter than the iMac Pro or 2019 MBP 16. The peak RPM I saw on the M1 Max was over 3000, but it took about 5 min of very high load to reach that, and it was still quieter other high-performance laptops (and many desktops) on the same task.
A complicating factor is the M1 Max MBP 16 has three different power/efficiency modes, each of which apparently has a different fan rpm curve over time. That is in System Preferences>Battery>Power Adapter>Energy Mode, which applies even if on AC power. In the low power mode it is quieter and slower to spin up, but at a performance cost. In the default auto mode it's louder under sustained load but faster. In the high power mode you can hear it more quickly, but it's still much quieter than my iMac Pro or 2019 MBP 16.
I ran three brief Resolve Studio 17.4.2 tests with a 60 sec 6k/24 BRAW timeline using various effects. Exporting a non-cached timeline to 4k/24 ProRes 422, the export times in each power mode were:
Efficiency: 7 min 28 sec
Auto: 6 min 20 sec
High Power: 6 min 20 sec
iMac Pro: 5 min 53 sec
This task was GPU-dominated. On the CPU side it mostly used the two efficiency cores. Workloads which heavily use both CPU and GPU might show different behavior.
I don't have numbers for the same task on other machines but the 2019 x86 MBP 16 would probably be slower yet much louder.
In low power mode the M1 Max seems to cap fan speed at about 2000 rpm, at which point it's still fairly quiet. I'm not surprised many people describe it as utterly quiet because it's hard to hear if blanked by normal sounds from HVAC, external drives, etc. If in low power mode it is extremely quiet. The sound does not emanate evenly in all directions. Just because you can hear it in front of the keyboard doesn't mean someone could hear it several feet behind the screen.
Other 3rd party tests seem to indicate the M1 Pro 16 might be even quieter and more energy efficient, yet not give up much performance if using the 24-core GPU.
