I too was experiencing this error on an Asus Dark Hero VIII and a 5950x. Memory is 32GB Trident Z Neo 3600mhz CL14 memory at 1.45V (Advertised XMP). The problem appears to be linked to a number of voltage settings and automatic power states in the bios. I could not activate the advertised speeds on my memory without constant reboots and BSOD/WHEA errors. My windows was also corrupted about 5 times. At first, I had to turn off CPB and PBO for my system to allow me to use the DCOP/XMP profile on my memory and become stable. However, I found a way around this allowing me to get the full potential out of my CPU and memory. These are the steps I took to make the system relatively stable after spending 2 days reading forums. This solution has also worked for others:
I hope this helps and please pass this on to any motherboard manufacturer/AMD technicians you are currently talking to. Thank you.
Loot at my solution...
60°C max peak in games
75°C max peak with Cinebench/AIDA64
You will lose 5% on single core benchmarks but there is no impact for games (same score with 3Dmark) and, especially your PC won't crash.
Definitely appears the voltage settings in the vendor's BIOS updates are stuffed and that's why the 'auto' settings keep crashing computers. Hopefully, in a few more months they will have the earlier adopter errors smoothed out.
I've this problem too, with a 5800x and a B550m TUF gaming.
In stress tests everything is ok, but on gaming i get a BSOD with
is it CPU fail? In this case RMA is the solution?