There have been some rather important updates to how the tech is *driven* though, but that's less noticeable on a Kindle than it is on other brands, because lab126 was always pretty damn good at that to begin with, and, to some extent, FW updates mattered on that front (e.g., FW 5.6.x saw a backport of a few fancy things introduced on the Voyage to all earlier devices running the same SoC).
I say "to some extent", because while some of the changes may be readily taken advantage of by the Kindle framework, some others weren't (or not until much, much later, or even delayed until newer hardware).
c.f., the
platform list for a breakdown of the history.
Some of those are definitely a SoC update (most notably Wario, then Zelda/Rex).
TL;DR: For your specific example, yes, there are definite advantages from switching from a PW2 (Wario) to a PW4 (Rex). (And that's coming from someone who used a PW2 for a good long while, and whose daily driver is now a Kobo Forma, which uses the same SoC as a PW4).
As far as the Oasis 2/3 (Zelda) is concerned, I can't really say how much the dual-core helps, as I've never used one myself, but what I can say is that, from the few platforms I'm familiar with (Kindle/Kobo, to a lesser extent PB/rM), the Kindle is by and large the one where a fancier CPU would make the most difference, given how the framework is built (highly modular, a metric shitton of running threads). But it'd mostly matter in non-critical use-cases, though. Reading? Eh, not so much, it's already pretty damn fast on Kindle.