Hi,
I can confirm I have this problem aswell on a brand new build. I have the same Motherboard and CPU, and also a 3090. My config:
CPU: 12900K
RAM: Corsair Vengence DDR5 5200Mhz 2x16GB (CMK32GX5M2B5200C38) [installed in A2 and B2 DIMM slots]
Motherboard: ASUS ROG Maximus Z690 Extreme
GPU: EVGA 3090 Kingpin [installed in PCIEX16(G5)_1 slot]
PSU: EVGA SuperNova 1600W
HardDrive: 1TB Samsung 980 PRO NVMe [installed in M.2_2] and 2TB Samsung 980 PRO NVMe [installed in M.2_3]
OS: Windows 10 Pro 64bit installed on C:\ Games installed on D:\

Symptoms: All games crash to desktop with no error after 5-20 minutes of playing. Halo Infinite crashes the fastest.
I've tested GTA5, Halo Infinite, Battlefield 4.

Windows error reports: I get 2 different faulting modules logged in the event viewer:
1. Nvidia Driver Crash
Faulting application name: HaloInfinite.exe, version: 6.10020.19048.0, time stamp: 0x61b2b778
Faulting module name: nvwgf2umx.dll, version: 30.0.14.7212, time stamp: 0x613fdb99
Exception code: 0xc0000005
Fault offset: 0x000000000020296d
Faulting process id: 0x1e50
Faulting application start time: 0x01d808f89c1d169d
Faulting application path: D:\Program Files (x86)\SteamLibrary\steamapps\common\Halo Infinite\HaloInfinite.exe
Faulting module path: C:\WINDOWS\System32\DriverStore\FileRepository\nv_dispi.inf_amd64_9be48e12ebceea24\nvwgf2umx.dll
Report Id: ba1c0a2b-a66f-464a-8cf8-2f83aa7b963e

2. Windows Kernal DLL Crash:
Faulting application name: HaloInfinite.exe, version: 6.10020.19048.0, time stamp: 0x61b2b778
Faulting module name: KERNELBASE.dll, version: 10.0.19041.1387, time stamp: 0x0b9a844a
Exception code: 0xe343000a
Fault offset: 0x0000000000034f69
Faulting process id: 0x8c4
Faulting application start time: 0x01d808f3bf54581c
Faulting application path: D:\Program Files (x86)\SteamLibrary\steamapps\common\Halo Infinite\HaloInfinite.exe
Faulting module path: C:\WINDOWS\System32\KERNELBASE.dll
Report Id: 272e840f-0821-4f67-b035-a8bd91986c36

BIOS settings are all default or auto for EVERYTHING except CPU fan set to "turbo" profile. No overclock, no XMP
BIOS version is now 0811 (from Jan 5th 2022)
CPU temperatures are at 35-45C while playing games and nothing over 80C while stress testing.

What I have tried so far:

1. Updating BIOS: this was the first thing I did, this actually did help. Before updating the bios to 0811, Games would crash almost instantly and I couldn't get through a single run of 3DMark Time Spy. After updating BIOS I can now pass 3DMark stress tests and IntelBurnTest.

2. Memtest86: Tested both RAM sticks and with each stick installed in A2 slot individually, they all pass with no errors every time.

3. Swapping ram to a different brand: I installed a G.Skill 6000Mhz ddr5 2x16GB C36 kit. No change.

4. ROG RealBench Stress test (this fails with instability sometimes) no result is generated but windows event view shows:
Faulting application name: LuxMark.exe, version: 0.0.0.0, time stamp: 0x55e59579
Faulting module name: MSVCR120.dll, version: 12.0.21005.1, time stamp: 0x524f83ff
Exception code: 0xc0000409
Fault offset: 0x0000000000074a30
Faulting process id: 0x2f50
Faulting application start time: 0x01d808e3577f8b07
Faulting application path: C:\Users\User\Desktop\RealBench\luxmark\LuxMark.exe
Faulting module path: C:\Users\User\Desktop\RealBench\luxmark\MSVCR120.dll
Report Id: 57573209-bda6-40df-9c08-2477d796935d

5. Increasing memory voltages: No change.

6. Re-install OS: No change.

7. Downgrading GPU Driver to v472.12 this helped, for a bit, and then I tweaked the graphics settings and now I'm crashing again. I can't fix it by rolling back changes to graphics settings, it still crashes.

I am at my wit's end. I'm going to try swapping out my GPU with a spare and I have ordered a new Power Supply to test. If it's not either of those it's maye the CPU or the Motherboard. This is getting really expensive and there are still shortages of everything. Please help me.

Hi OldManDan,
Do you also get "Nvlddmkm stopped responding but recovered" error when games crashed in event viewer?
May I have the following information?
- the serial number of your motherboard via PM
- Os version and OS build
- the version of the crashing games
Thank you.

Hi STARRAIN,

I did see a crash mentioning nvlddmkm at least once but that was before I re-installed my OS. I dont see that particular error anymore.

Windows 10 Version: 21H2
Windows 10 Build: 19044.1466

The game that I am testing and is crashing is Halo Infinite
Version: 6.10020.19048.0
App ID: 1240440
Build ID: 7861287

I have a progress update on this.

After updating to the latest NVidia driver, I tried to play and actually was able to play for 30 min without a crash. I then changed the graphics settings and restarted the game. The game then crashed repeatedly on level loading. (not during gameplay like usual, but at the main menu while loading a level) I could not fix it by reverting the changes made to the graphics options.
I then deleted 2 folders:
C:\Users\User\AppData\Local\NVIDIA
C:\Users\User\AppData\Local\HaloInfinite

When I restarted the game afterwards I was able to load the level again, and did not crash for at least 25 minutes.
Is this a hard drive problem? Were cached files being corrupted?

I was then following another thread about WHEA errors related to the NVMe storage drives. This thread was also discussing random crashes.
https://rog.asus.com/forum/showthread.php?126865-Hero-Z690-having-issues-with-getting-windows-errors...

I checked my event viewer and saw similar errors so I disabled PCI Express power management. After doing so, my games still crashed.
Then I thought if the hard drive wasnt working right, maybe windows and the drivers did not install correctly?

So I completely re-installed windows after formatting the drives. I have not had a crash since, and have been able to play for at least 70 minutes straight. It looks like this "Power-management" feature is starving the drives for power and causing a cascade of issues that look like RAM, CPU, or GPU problems.

I'll report back if I get more crashes but so far so good.

Hi OldManDan,
Thank you for your reply.
Is PCI Express power management enabled or disabled after format the drives and completely reinstall windows?
other bios settings are all default except CPU fan set to "turbo" profile?
Thank you.

Hi OldManDan,
If you don't use M.2_1 slot, it is okay to install the graphics card in PCIX16(G5)_1 or PCIX16(G5)_2, but we recommend to install in PCIX16(G5)_1.
Thank you.

I have the excact same issue with the same errors in event viewer.

i9-12900K
Team Group DDR5 6200Mhz
NVIDIA RTX 3090 Founders Edition
Asus Z690 Apex
Samsung 980 PRO 2TB (OS Drive) 980 1TB , 970 EVO Plus 2TB and a 970 EVO 1TB drive. So Four NVMe drives.
EVGA G2 1600W
Windows 11 Pro

Have not played many games since i build it late december, but Fortnite and CS:GO is not crashing to desktop, but BF4 is, sometimes i can play for a while, other times not.

No bsod, it either just exit/crash to desktop, or it will freeze for a few seconds before im back at desktop. At Battlelog it says i'm still ingame and playing.

Faulting application name: bf4.exe, version: 1.8.2.48475, time stamp: 0x58ab1cc0
Faulting module name: bf4.exe, version: 1.8.2.48475, time stamp: 0x58ab1cc0
Exception code: 0xc0000005
Fault offset: 0x00000000008cc28f
Faulting process id: 0x413c
Faulting application start time: 0x01d86de2c2964fba
Faulting application path: C:\Program Files (x86)\Origin Games\Battlefield 4\bf4.exe
Faulting module path: C:\Program Files (x86)\Origin Games\Battlefield 4\bf4.exe
Report Id: 33e63401-ada4-43e2-80e2-2564a4a656f2
Faulting package full name:
Faulting package-relative application ID:

Faulting application name: bf4.exe, version: 1.8.2.48475, time stamp: 0x58ab1cc0
Faulting module name: nvwgf2umx.dll, version: 30.0.15.1277, time stamp: 0x6272de84
Exception code: 0xc0000005
Fault offset: 0x0000000000b89076
Faulting process id: 0x6798
Faulting application start time: 0x01d86ddddc654275
Faulting application path: C:\Program Files (x86)\Origin Games\Battlefield 4\bf4.exe
Faulting module path: C:\Windows\System32\DriverStore\FileRepository\nv_dispig.inf_amd64_d5d5b9f929f4cb65\nvwgf2umx.dll
Report Id: f01329b4-2797-469e-8598-08076b93c54f
Faulting package full name:
Faulting package-relative application ID: