I would also disagree with you on performance gain for higher clocked memory. Ryzen 3000 has a noticable uplift in performance especially in gaming the higher you crank infinity fabric cuppled 1-1 with frequency. Gamers nexus tested this. Memory timings make very minimal change but frequency and fabric shows a huge difference anywhere up to 7%. Gamers nexus actually recommend stock cpu frequency and concentrate on memory for overclocking as it makes a much larger difference. On bios F12F i can run rock solid stable at 3600mhz 1800 fabric tight cl18. On bios f20 I can't even run 3200mhz stable and drop betweem 7 and 12fps. For me that is the difference between running 1080p and 1440p over 120+ fps for decent 120-144hz competitive shooters like cod warzone. If my choice is game in 1080p or buy a new board.... easy choice. These problems have now been here over a year through 14+ bios. A year and still mass users unstable. Bare in mind a lot of us will buy a ryzen 4000 and will not want to risk another year of poor performing bios updates. We are likely to buy a non gigabyte x590 when they release. Gigabyte list the x590 in the x570 bios after F11. So we know the boards really are coming.
Gamers nexus tested memory kits from under 3000 to 4000mhz. In GTA for example There was a 12% increase to performance between 3200mhz and 3866MHz. What are you basing no performance gain on? An opinion? Or a controlled test?
CORSAIR VENGEANCE LPX 3200mhz . Typhoon confirms samsung B-die. Like I said as common as it gets and has been listed as supported since the boards release. I can't give you the full model code cause at work.
What make and model of RAM is it and is it on the QVL list?.. being common means nothing.
As further back in the thread I'd not worry about under clocking RAM as even pretty large under clocks make tiny differences in PC performance. I gave up years ago fine tweaking RAM to it's absolute "Maximum" overclock performance as it's a waste of time unless your going for "competitive" overclocking... the rest is just wasted time. I ramp it up towards it stock max and get to a high stable point then step it back down.. currently running 3600 RAM that seemed rock solid at 3600 but I still reduced it down to 3400.. more stability (even a small reduction in MHz is a good stability improvement) and insignificant performance hit.
Personally I think the general lack of XMP stability comes from it not being Intel which is more supported due to the dominance over the years (that's changing now). Though saying this I have numerous RAM modules I can get stable at stock XMP on X570 I can't on on Intel boards of varying Brands.. so it's not just X570 nor Gigabyte boards.
Have you compared the XMP timing settings values between each Bios? I'm guessing they may be more aggressive and if it makes people feel better seeing a larger MHz number for their RAM then slacken the timings and up the MHz.. ..
My personal take in 2020 is I buy RAM and expect to run it lower than the sticker Spec.. I'll find RAM at a good price/MHz/Timings and check the QVL list.. if it's not there I may just risk it & buy it anyway if it's a good deal.. If I wanted to genuinely run at 3600 for some reason I'd buy 3800 or 4000 anything >3600 is a Bonus and ensure it was on the QVL to make totally sure but in the silicon lottery it's still not guaranteed.