I pretty much agree with the sentiments of @MattiasNYC and partially with @raino
In my experience, CPU core speed seems to come into play especially for
- modelled instruments
- intense serial audio plugin chains (and this can also get worse when routing through multiple layers of group tracks with audio insert FX, which can creat a longer serial processing chain).
When rendering one of my recent projects (final mix to stereo), it was interesting to observe that 1 thread was pretty topped out at close to 100%, while half of the others were mid level loaded and the remaining half was sitting closer to idle.
The project had a pretty intense guitar track fx chain (approximately 8 inserts) and the master stereo out had a relatively average mastering fx chain…
The VST instruments were mostly sample based, so not much CPU stress from those tracks. But the orchestral tracks were submixed into orchestral sections (strings, winds, etc.) and that ended up going into an orchestral submix, and that into an instrumental submix, before hitting the master stereo out. So that string of submixes forced a bunch of stuff into sequential, rather than parallel processing.
For my recent system build I ended up going for a Ryzen 9 7900x, giving up a little bit of top end Intel MHz for having lots of equally powerful cores (i.e no efficiency cores).
But I gave up a few AMD cores to get to the generational top end of AMD core speed. – So I guess, I ended up somewhere between the top speed of Intel CPUs and the top core power of AMD CPU’s.