A chain is only as strong as its weakest link.   A processing system is only as fast as its slowest component, multiplied by the queue-size that results from that differential in speed.

It is very, very often the case that “the slowest component” is precisely the same as it has always been:   the disk drives.   Not the networks, not the CPUs or the number thereof.

In my opinion, the best way to get definitive answers is:   simulation.   Except in the very rare case of processing that actually involves CPU-intensive activity, the number of CPUs or cores can almost always be omitted.   What matters are things like ... cache hit ratios, data distributions, and I/O avoidance.   And this has more or less been true for about sixty years now.