Build: i12900k w EVGA 360mm AIO, 3090 Ti FTW3 Ultra w EVGA "Powerlink 41s" power adapter, EVGA Z690 Classified, G Skill DDR5 6400 16gbx2, EVGA 1300W G+, Samsung 980 Pro 2TB, 5xNoctua 140mm industrials. ROG StrixXG27UQR monitor using its DP 1.4 cable, case is Fractal 2 Meshify XL (planning on doing more for the build with a custom waterloop in due time)
 
Context: I run 4k and frame cap at 60. Randomly when I'm gaming, my GPU will literally shut off. I usually hard restart my computer at that point since I literally can't do anything. Absolutely no output. You can hear a prompt when this happens (headset), and the computer is still running just fine in the background, but I made the mistake of hitting enter since it's obviously an error code. That lead me down a troubling path that only unseating and reseating the GPU would fix, along with clear CMOS several times. Thought I bricked my **** at first. I haven't even used this card for a month. Honestly have had this issue from the start and I've been gimping with it.
 
I'll also add that my CPU temps haven't gone over 50C and GPU temps are 60C and below, generally around 50C.
 
Even running default BIOS, NO overclocking whatsoever not even XMP, stopped undervolting CPU, DDU'd and clean installed the studio driver instead since I heard they're more stable, fast boot off, most of my drivers are all up to date especially BIOS and graphics, or, at worst, a month out of date.
 
I also have to restart my computer upon boot half the time since derp-a-derp DP signal thing didn't handshake my monitor or whatever you guys call it, but I figure that's a separate issue at this point.
 
I've been looking looking through forums and posts, seen similar things, but not to the extent of mine. Generally it's people complaining about dual displays, which I'm running only 1. I've seen other people's post suggest it's a DDR5 issue, I've seen others suggest it's a power issue, but what really scared me is the possibility that the graphics card itself is bunked up. Saw some of that too, then RMA'ing only to get a refurbished that runs hotter than the original GPU. Kinda made me wonder if there's an Nvidia F YOU EVGA conspiracy or something lollll just a thought based on a skewed bias
 
And if it matters, all 3 Displayport ports work fine. When one isn't emitting signal, none of them are. Haven't cared to try HDMI since my monitor doesn't have speakers, though if that's a temporary solution, I can try it.
 
And about the power hookup to the GPU, like I have stated above, I'm using EVGA's Powerlink 41s for the 3090 Ti, NOT DAISY CHAINING, which, excuse me for not knowing the jargon, has 4 8 pin VGA plugs connected to it and 4 VGA slots on the PSU. It shouldn't have any issue getting the power from the power supply. Weirdly enough, when speculating the power adapter port that plugs directly to the 3090 ti, (I think it's a 10 prong power port?) the top 5 prongs are bronzed while the 5 are shiny, suggesting perhaps an overload? But if I recall right, it looked like that upon a fresh install so I shrugged it off. No funny smells.
As for the outlet, it's only my PC and monitor connected directly to the outlet. This room's circuit is a loaded (gateway, average minifridge, a couple peripherals) but nothing drastic, not that I can honestly change it.
Some of you might call me out for having 5 Noctua 140mm industrial fans. Yeah, they're heavy draw at 6.6W, 12V each. [They Spank. My 2 year old told me, "Hey daddy! It's the vaccuum!" on startup.. facepalm. Honestly, I don't want them connected to the montherboard and I've searched for a solution to no avail. Maybe because I don't know a technical term.]
 
Suggestions? Tips? Ridicule? I'm relatively new to custom building, however, I'm pretty thorough, so keep that in mind. I'll reply when I can, so please have patience.

I made a build for a good friend identical to mine (same board, CPU cooler, CPU, case) except he has 2x32GB RAM with a base 2080. I was hoping I could be lazy enough to skip part swapping, but I think that's a damn good start at this point. I'll swap him GPUs for a time and see what his experiences are.
 
I'll swap tomorrow when I get off work. I'll give updates. I usually encounter an issue within 2 hours of gaming on my computer.
 
I appreciate your feedback in a timely manner. In the meantime, any other thoughts and opinions are appreciated as well.

I personally haven't tried using other power source means.
 
Update: My bud and I swapped GPUs. His setup is exactly the same as mine, except he has an audio card and a USB expansion card hooked up with plenty of MIDIs (musician). Anyway, he ran the 3090 ti without the powerlink at first. No issues whatsoever after 2 furmark tests and an hour or 2 on GoW5. Then I had him use the powerlink. He hasn't used it long, but there seems to be no issues. He hasn't stressed it much for too long, so to be continued on that front. I have his 2080 in my PC and I've played 3 hours of Destiny 2 and ran Furmark a couple times no issues. I suppose I should technically note I'm on a different outlet for a different room hooked up to my Samsung 65" instead.. He'll do a couple more days of testing under different conditions, then we'll swap back so I can test things more particularly in my environment, like different power connections.
 
All in all, the 2080 brings absolute stability to my PC. This all leads me to believe it's a power issue on my side as the 3090 ti draws more than double the wattage as the 2080. 2080 is around 210 watts and 3090 ti was around 450 watts. My power supply is new, RMA'd from EVGA as the first one **** out on me for some reason. I've been thinking about upgrading to a different PSU, there might be issues with this line. The problems are similar, but not the same when comparing my issues to the other PSU before RMA. That was also with a whole different setup.
 
Whatever.
 
Any comments or ideas?

^^^^ That
Also, what psu did your buddy have....losing video signal consistently while loading the GPU (gaming) I would consider 2 possible things:
1. PSU
2. RAM (however, you said you ran system at stock w reset in BIOS), however, loading defaults sometimes still requires a bump in RAM voltages (within BIOS ), but generally this is only when enabling XMP.
Perhaps your "buddy" would let you try out his psu, (just offer up some cable management time) ;)

Final update:
 
Changed the MOBO battery, but that wasn't the issue.
When I used my buddy's 2080, I had to change the location of my tower to my 65" TV since I lent him my 4k monitor (his is HD) to produce identical results and strain. Ever since I used a different outlet, I've had no issues.
Now that we've swapped GPUs back, I kept my tower in the same place just to test it as I was becoming skeptical, and sure enough I'm having literally 0 problems. Did a 20 minute CPU overclocked stress test on Cinebench, 6400 XMP profile on the RAM, binge gaming in 4k and I haven't ran into a single issue whereas I normally would.
 
You'd think an outlet issue would appear sooner or be more obvious. We've updated outlets to code in the past couple years and haven't had any issues in this room, but a PC that takes this much wattage is bound to find issues pushing past normal load on the circuit. I wonder if mice chewed some wires. After checking some basics, voltage, amperage, I'll call an electrician.
 
I shouldn't have doubted an EVGA PSU, that was my bad. Too pessimistic these days.