My Classified Z790 system, (BIOS 1.11, Intel 13900KS CPU, 4x32gb Kingston Fury DDR5 RAM DIMMs, NVIDIA 4090 GPU, and Windows 11 with all updates applied) has been my daily driver since last fall. I did have a mild undervolt configured in the BIOS, but once I discovered this issue, I reset all BIOS settings to default, and was able to recreate the problem.
 
Basically my system now, will always crash hard, while playing most modern games. I'm primarily playing Fallout 76, but I also confirmed this with Starfield as well. After playing the game normally for a few minutes. The video will cut out, I will no longer hear any game sounds, and nothing will let me recover the system, unless I power it off and on again. I've not noticed it happening when I'm just using the system for regular task, web browsing, email, etc. Just when playing games.
 
My initial RAM was 4x32gb G.Skill Trident DDR5 DIMMs, and I noticed the initial POST codes when the system crashed as being memory related, displaying 45 then 51, and 46 and then 52. I had noticed during the original installation, that even though there was a single XMP Profile for the G.Skill DIMMs, if I tried to enable it, the system wouldn't boot, (so I always had the setting as Automatic).
 
As a result, I assumed this problem was being caused by faulty memory, so I replaced them with Kingston Fury DDR5 DIMMs which do show two XMP profiles in the system BIOS, and the system will boot normally if I select the XMP profile. (However, I decided not to continue testing with any XMP enabled, and simply reset the BIOS to system defaults for all of the current testing)
 
Now, when the system crashes I am greeted with a number of CPU related POST codes, such as 56, 57 , and 58. Interestingly I noticed a similar post in the Intel forums where a user was describing a similar situation with an Asus motherboard.
 
https://community.intel.com/t5/Processors/brand-new-i9-13900k-CPU-mismatch-self-test-microcode-Q-code/m-p/1430027
 
This is a water cooled build and so replacing the CPU means disassembling the loop, which I would like to avoid. I've created a service ticket to ask EVGA support for help, but if anyone else has seen this, or similar, issues, and has a solution, please feel free to comment. Otherwise I'm curious about anyone's experience using POST codes to determine system issues. If you have any general advice/tips/etc., please feel free to share those as well.
 
TIA..

Step one is to test the system at 100% stock and only with 2 dimms installed. If everything is working fine after a week of hard testing, I mean everything. Then install the other two dimms.

If you aren’t willing to break your system down to swap hardware you’re gonna have a hard time fixing issues that come up. Since you don’t want to break your system down the only thing I could suggest would be to try increasing the voltages of your cpu a bit if you continue to have issues at stock settings. Increasing voltages may help you get stable if it’s really cpu related. In the end you may have to swap hardware anyways to figure out the issue.

You really need good troubleshooting and sometimes that involves breaking down your system. I know it's a pain but sometimes there is no other choice. 

Some of those issues can be created by bad CPU pressure. I had a whole myriad of random issues when I was cranking down the water block too tight. I have an Optimus block and it doesn't have a "bottom-out" type of mount with spring retention, so you just sort of have to do it by feel and be really strict about cross-tightening so that it applies pressure evenly. I was getting situations where it would boot fine, then restarts were bad. Or the cold boots would give a post code, but restarting would be fine...it seemed like everything changed as the system heated up a little...like things were moving or something. I think I was causing flexing of things, which resulted in bad pin-contact under the CPU because I wanted a really good cold-block mount. Turns out performance is great and I don't get weird nonsense if I just mount it more gently lol. Also, the after-market LGA-brackets were making that particular situation worse, but mainly because it was really hard to get an even mount with the system upright until I bought a bracket for behind the motherboard. This added a lot more stability to the mount and let me be much more careful in how tight I was cranking things down (and how repeatable, etc.).
 
You also could have a damaged CPU socket, which could be making pretty good contact...sometimes...but not all the time, resulting in the quirky behavior.
 
I'd start by making sure BIOS is up to date and checking the CPU socket and mounting pressure.

Apparently the issue I've been having since February with my system instability, (and reported to EVGA technical support), is now recognized and documented as being an Intel CPU BIOS issue discovered by NVIDIA. The affected CPUs are Intel 13th and 14th Gen CPUs, and it appears to affect the high end SKUs specifically.
 
Intel has created a thread in their community forums for discussing this issue here;
https://community.intel.com/t5/Processors/Regarding-Reports-of-13th-14th-Gen-Unlocked-Desktop-Users/td-p/1575863?profile.language=en
 
And Rad Game Tools created a blog post about the issue here;
https://www.radgametools.com/oodleintel.htm
 
In those threads various folks have commented on several different BIOS settings which they have adjusted to prevent these system crashes. So if anyone has diagnosed this and discovered the proper EVGA BIOS settings that work, please feel free to respond to the thread.
 
TIA..

I'm running 13900KS CPU with Z790 Classified, zero problems. I don't overclock, I did limit CPU power usage to 320W per intel spec for 13900KS in BIOS.
You can find those settings under:
ADVANCED --> CPU Configuration --> CPU Power Limit (set to 320)

 
Default is set "unlimited" to: 4096W

Thanks very much for sharing that information.
 
I just now went into my BIOS and changed the CPU Power limit to 320 as well. I will now monitor the system and see if these crashes have been mitigated.

Unfortunately, I've not been able to stop the random system crashes caused by this issue yet by setting the CPU Power limit to 320. The Intel recommendations for addressing this issue are posted at the top of this Reddit thread;
 
https://www.reddit.com/r/intel/comments/1diylqo/official_intel_guidance_for_13th14th_gen_power/
 
I'm currently running the beta 1.13 BIOS for my Z790 Classified motherboard, and have tried a series of different power settings to try and mitigate this issue, based on these Intel recommendations, but I've not been able fix my system crashing. For example, setting the CPU Power Limit in the EVGA BIOS appears to set the PL1 and PL2 limits. The Intel recommendations mentions setting an "ICCMAX Unlimited" setting for the CPU as well, and the settings for my 13900KS should be;
 
ICCMAX     400A
PL1           320W
PL2           320W
 
I just don't know how to set the ICCMAX separately from the PL1 and PL2 limits in the EVGA BIOS. When I contacted EVGA support they claimed that they were not aware of any issues like this, and unfortunately one of the articles on the subject;
 
https://videocardz.com/newz/intel-releases-etvb-microcode-fix-and-new-default-settings-for-13-14th-gen-core-i5-i7-cpus-instability-investigation-still-ongoing
 
states that a fix is being provided by Intel to the motherboard manufacturers, which would require a BIOS update. Unfortunately I'm not confident that an updated BIOS which incorporates the stated fixes will be available any time soon from EVGA.
 
If anyone can parse the Intel instructions into EVGA BIOS settings, I'm willing to try any recommendations.
 
TIA..

Can you try reducing your max p core frequency by 200Mhz, this will only affect your 2 prefered 6ghz cores and keep your all core frequency the same. This resolved my crashing issue on my 13900k. That last 2 VIDs are just on 2 cores so you won't notice a performance difference in anything, you just won't get to see that gorious 6Ghz anymore lol.