GPU Constant crashes

xDaTenshii · May 6, 2023

So this is something I have been dealing with since I basically purchased the GPU;

Current Setup :
Ryzen 9 7900x
ASRock 670E Steel Legend
G.Skill Trident Z5 6000 DDR5 CL30
Gigabyte Aorus RTX 3090 Master rev.1
EVGA 1000w supernova G6
Corsair Cappellix h150i Elite

Problem : Playing games or really anything; or even any multitasking randomly crashes; fans ram to 100; Everything in the background continues to run except will disconnect me from games or whatever but sound will still play from say a video;

What I have tried :
Basically swapped every component of the pc including :
Ram; PSU; CPU; Mobo; and Cooler.
Reinstalling any and all drivers, as well as using new M.2 drives and reinstalling windows.
Have even sent in the GPU for a 3 week RMA process where they concluded basically Nothing was wrong with the gpu, (being the only part i havent switched) insisting that it to be swapped, they eventually did, however they kept the same rev.1 model.

Old setup i swapped from :
Ryzen 3600
Gigabyte 570x wifi elite
Corsair Vengeance 3600 DDR4
Same GPU
Gigabyte 850w PSU cant remember the exact model (it was the one that was exploding) which i got the recall for after hand however never had the box + such when they did that and was planning to replace it anyhow.

The problem has been occuring since i first got the GPU, it was easily repeatable before (atleast for me) as certain games simply would not load or play; games such as Halo 1 or 2 would get past the initial loading screens and intro and within the first few minutes of game play it would crash almost the same part every time. Magicite I could open up the game and the moment i clicked multiplayer it would load up the UI for it and Immediately crash. Other games would crash Out of the blue it seemed; another one was Borderlands 2 or prequel it would ONLY crash during the Character selection when creating a NEW character after it gets through the intro, pressing the button fast enough to accept it and go through would progress, but if you stopped on that screen for more than a second it would crash. Gigabyte claims they tested it for like 17 hours and couldnt replicated the crash, though i seemed to do it no problem. Once RMA'ing the GPU I got back a swapped unit still Rev.1 and started to play games; At first it seemed ok and went a few weeks where i didnt see any crashing, maybe once or twice but nothing to noticable, as time carried on again the crashing occured more and more frequently until its back to DAILY and likely even more than daily sometimes playing games like Path of Exile currently it will crash just completely randomly; could be idling in hideout or actually running stuff; it has also (and i believe started or when i first noticed crashing again) was in Tarkov; as well as other games it will randomly crash for. I cannot seem for the life of me to be able to Consistently recreate, though I have noticed that when I am watching another Video or stream it seems to happen more often while I game.

Side note : I am using a Triple monitor setup with 2 1080p monitors on the side of a main 1440p monitor all 3 are Asus.

Possible Theories on what the problem could be :
1) The Rev.1 Version of the 3090 has only 2 8pin connections; I stated that during the RMA process they didnt move or send me the Rev.2 Version which has 3 8pin connectors like pretty much every other 3090.
2) Maybe a PCIE Glitch in their software? Someone else said they had success with a different gpu manually switching to Gen 3 (though im fairly certain i tried this already i will try it again)
3) I also thought that the possibility of something glitching the crashing from the display this particular GPU has; it seemed to happen more frequently once i changed the display to a custom gif; however taking that off and resetting it didnt fix or even slow down the issue.
4) Currently Trying to limit the Power on MSI Afterburner at 85% instead of going to 100 or letting it do whatever; I know i have tried this before and it temporarily helped but ultimately didnt; I have a suspicion that it has something to do with some sort of brief spike and the gpu not handling it and crashing?

Honestly I just dont know where else to turn or what else to check; I simply cannot send my gpu in AGAIN for another RMA process without having some sort of GPU in my system, especially when it takes them so long to even get it, nevermind looking at it. Especially when its likely going to result in them just telling me there is nothing wrong with it and sending it back again. (they also arent supporting the buy another one and send them mine and then they refund or whatever anymore apparently)

Literally my last effort to do is ask on here for any kind of help in this regard; I really feel like I have tried testing everything I can and without just buying another gpu which I cant exactly just afford another couple thousand for another GPU or even another few hundred, it has taken me years to save up for all the components I currently have and has been a long term goal to have a PC like this. Also everything is in CAN Prices which feels like its absurdly more expensive then the rest of the places like US lol. Thanks for any ideas or help in this manner.

csmt · May 6, 2023

Do you have another psu to test?

Also I had these issues with ny card where it would crash and fans 100%. I solved it by renewing the thermalpaste and I suspect the vrm cooling pads werent in the correct place.