I own a 8700K (gaming rig) and R5 3600 (ITX lounge PC), I decided to drop my 5700XT into the Ryzen system just to see if there is much difference in gaming performance at this level of GPU, since most reviews test with 2080 Tis to reduce GPU bottlenecks.
In all honesty, there is not a noticeable difference that I can discern with the sole exception of Far Cry New Dawn, which runs much better on Intel for some reason. I also game at 100Hz (well, 110Hz since I overclocked my monitor) but at UWFHD (2560 X 1080) instead of QHD so if anything you'll see even less difference as you'll be more GPU bound.
The difference is still measurable in terms of benchmarking, but it's small enough that I probably wouldn't be able to tell in a blind test which system is which.
That being said, I dont have a specific requirement of 97+ fps at all times, and I tend to max out my IQ settings unless I'm gaming competitively online, at which point I'll drop settings down a notch to maintain consistent 100fps+ performance.
If 97fps+ is the metric to go by, then a 9900K would generally provide the higher min fps compared to the 3900X - assuming you are gaming with a suitably fast GPU and / or more CPU bound settings - NOT the ultra settings that were linked earlier showing all the CPUs bunching up due to GPU bottlenecks.
Ultimately, if your end goal is consistent 100fps+ then you don't game at ultra settings, period. Even a 2080 Ti won't guarantee that at all times, let alone a 5700XT.
'High' settings usually show a 20 - 30% uplift in fps compared to 'ultra' without a huge loss to IQ, and in that scenario a 9900K could be the difference between say, 90fps mins and 100fps mins.