I have been fighting stability issues when playing BF2042 and New World the past couple of months.  Key hardware summary:
 
Thermaltake 850W TPG RGB 80+G FM ATX PSU
X570 AORUS PRO WIFI
AMD Ryzen 7 5800X
EVGA 2080 Ti XC Ultra (BIOS Version: 90.02.0B.40.83)
 
The symptom is that the displays will stop updating, they basically freeze with whatever image was on the screen when the problem occurs.  No keyboard input is possible (CTRL+ALT+DEL, WIN+CTRL+SHIFT+B), yet I can see numlock/capslock light toggles if I press those keys.  Usually if there is music playing in the game, that continues to play indefinitely.  The only way to recover is to hold the power button on the system, forcing a full power cycle.
 
I also play Overwatch pretty regularly, but I've never had similar issues happen when playing that game.
 
I'm running the latest software and BIOS versions all around.  I have been using a moderate GPU overclock, set using OC Scanner in MS Afterburner for quite a while.  I noticed that the problem was affected in a positive way by resetting to stock settings, but I still ran into the occasional voltage drop below 11.4V and eventual hang in game.  I found that I could leave the overclocking in place and reduce the Core and Power Limits to stock and I would get similar results, meaning better stability but still encounter the occasional crash.
 
I won't go through all of the crazy troubleshooting that I've done, but I think I've narrowed it down to perhaps the GPU not getting enough power from the mobo and PSU.  I came to that conclusion by watching the voltage on the PCIe +12V, 8-pin, and FBVDD Input Voltage readings in HWiNFO.  What I observed in both of the problem games is that 8-pin Input Voltage was often dropping below 11.4V, which from what I've read, is out of spec.  There is another reading for GPU FBVDD Input Voltage that is also dropping below 11.4V, which I also assume is bad... but I've found limited info on where that power is derived from and if it is a red herring for this problem.
 
I have observed that stock settings clearly reduce the voltage drops observed in HWiNFO.  At this point, I've settled on a Power Limit of 92%; I will have to do more testing to see if it really is stable there.
 
Sorry for the length of this post, but I'm curious if anybody has an opinion on likely root cause here?  At this point I'm thinking I need to change the PSU cables or the whole PSU itself.

HeavyHemi
 
During your trouble shooting did you reseat the GPU and PCIe power connections? Try a different PCIe cable? You should not be seeing a drop that large. Under full load I rarely drop under 11.8 and I'm pulling over 300 watts on my 1080 Ti. Are you using separate PCIe cables for each connection?
If reducing the GPU load and or reducing the GPU clocks results in higher stability, there's really only one likely culprit: your GPU is degrading. It might be worthwhile to change the TIM as a void or hotspot may have developed that doesn't show in monitoring. The secondary culprit could be your PSU. I'd try to resolve that voltage drop.

 
Thanks for the reply.  Yes, I did remove and reinstall the GPU and PCIe power connectors -- at least on the GPU side.  Further fiddling is more of a weekend task that I haven't bitten off yet.  I ordered replacement cables for the PSU; I have extensions for the cables that are visible inside the case, and I am wondering if that is interfering with power delivery.  Thermaltake has some funky proprietary connectors on their PSU's, so I ordered full replacement cables from CableMods, just waiting for delivery.
 
Also, I have encountered more hangs with the 92% Power Limit, so still haven't found the right combination to get things stable.  I'm skeptical that this is a software problem, but I guess I should use DDU and install clean drivers too.