Hi
I think you should read the information regarding the Zen architecture and infinity fabric as you seem to not quite understand the points being made to you.
https://en.wikichip.org/wiki/amd/infinity_fabric
https://en.wikichip.org/wiki/amd/microarchitectures/zen
I feel that your reasoning as for why this doesn't make sense are because your understand of the workings is a bit simplified.
Your point that the ram doesn't do processing is correct. From a traditional point of view RAM feeds data into the CPUs. Zen architecture uses CCX modules with 4 CPU cores per module, there are two CCX modules in every consumer zen CPU with a couple of exceptions for the Vega graphics versions. You will always be going across the infinity fabric to perform computation that requires more than a single core. To this end when you run anything intensive you will will likely be running a job across two cores that are on different CCX modules so shared information will be using the infinity fabric.
Infinity fabric runs at half the speed of your ram, so for 3000mhz ram it runs at 1500. With the speed higher you can feed data to the CPU cores quicker, a massive reason why increasing the ram speed can increase your benchmark numbers is because you can feed the data to the CPU faster for processing. Not because the ram processes the data. This concept has more importance on Zen architecture. On Intel CPUs all cores are within the same silicon so you don't get the same speed up.
https://www.google.co.uk/amp/s/www.eteknix.com/memory-speed-large-impact-ryzen-performance/amp/
If you read the pages linked here I don't imagine you will be confused anymore, but even if you are I would recommend you accept that you're wrong and keep reading, that's how we all learn.