GPU processing benefits from a faster GPU and GPU memory up to 6 gb. If you look at my graphs you will notice that GPU load dips low most of the time, so average load is only around 70-80%, so at some point a more expensive GPU will not help anymore. My RTX 2070 Super is *not* faster than my former RTX 2060 non Super when both use the same clock-rate. So the extra cores of the 2070 Super do not help anymore.
CPU processing benefits from more cores up to a point. On my 5 GHz 9900K GP uses 16 processing threads on 16 logical cores, since only 8 cores are real physical cores and the rest is Hyperthreading I get about 70-80% total CPU load. More cores might help, but I am already seeing single threads not running at full core load.
So at this point is seems that Topaz first needs to further optimize their code before it makes sense to throw more expensive hardware at the problem.