With such mixed gaming results in the reviews around the web it has to be firmware.
It could be different reviewers using different PL1 / PL2 settings with different Tau expiry and different benchmark lengths.
Even if two reviewers use Intels specified figures rather than the motherboard defaults if reviewer A does a 30s game benchmark and another does a 5 minute game benchmark the average FPS at 720p is going to be lower in the latter because that longer test goes past the full turbo mode and your average clockspeed for the run is lower.
On top of that RKL does well in some games and there is genuine improvement so if a reviewer has a game course that happens to include a lot of games where RKL does better than average you get a better result.
This means the variability in reviews can easily come down to different methodology and different games. I tend to prefer reviewers who very clearly explain their methodology so TPU, GN, HUB(TechSpot), Anand, ComputerBase are my goto reviewers.
Intel did not force you to use DDR4-3800 gear 2. Not guaranteed to work on either system. You decided to do that due to pressure from the AMD community.
This is what happens when you use 2x 16GB DDR4-3200 CL16 gear 1 on Z590 Rog Maximus Hero XII and 4 x 8GB DDR4-3200 CL 16 with X570 Aorus Master. Both systems are running memory synchronously.
It's actually far, far more likely that a typical DIY type will use this config.
View attachment 194722
3800 C16 is a pretty common and easy to achieve setup for both Intel and AMD, or at least was until Intel changed their IMC. Further this tester released videos of their benchmark scenes (really good idea) and if the benchmark is actually exactly as the videos play out most are sub 60 seconds so all take place in the turbo window at Intels PL1/PL2/Tau settings which will inflate the Intel numbers. They also use different motherboards for 10th gen and 11th gen Intels so it is not as like for like in the generational comparison as you could get. Also 2x 16GB vs 4x 8GB seems like an unnecessary difference when you could run 4x 8GB or 2x 16GB in both AMD and Intel rigs.
On top of this HUB and GN who use 4x 3200C14 ram had very different results with the 10900k being ahead of the 11900k in the HUB 10 game average.
Also with all the crud CapFrameX was spouting on Twitter and over at ComputerBase forums I would not trust their results as much as the likes of GN, HUB, TPU, AnandTech and ComputerBase.