Hello, in desperate need of some help.
A little backstory, I had bought a ROG Strix 6900 xt LC OC from eBay (yeah, I know) used since it was the last piece to the build and was getting a little impatient on finding this GPU. Found one gently used as they described, was less than 2-3 months old, 2 owners with the first owner not even taking it out of the box and direct selling to the second owner. He had used it for gaming, nothing more.
GPU received, no obvious issues at first glace. Card still had the white circle sticker over one of the cooler screws on the top of the GPU. I had hooked it up to a previous build that I was using while building the new one (5600x Ryzen/XFX 6800XT/RX850w PSU/280 AIO corsair/TUF 570 mobo/NVME x2) and using a corsair 5000d with 10 QL120 fans. Never had any heat issues, not even in the slightest. Multiple times after booting up FFXIV, Witcher 3, Escape from Tarkov, and numerous other games that are relatively AAA or have been around a bit games would black screen crash and computer would not be able to successfully re-boot while ending up hanging. Updated all drivers manually, latest BIOS, tested on both manual OC/auto OC/stock profile, tried all power settings, and multiple setting changes such as game mode, confirmed latest windows, etc. etc. DMP files have been assessed with no discernable cause, just getting the standard Kernel 41. DDU, put my old 6800xt and my XFX 5700 xt with no issues after fresh DDUs. PSU tested, no issues, no ram issues on Memtest. Stress tests for the most part had no issues, it was only when gaming. New install of windows, completely fresh with game crashes persisting no matter what games i play.
Built my new build with 5800x/Ryujin 360/Helios case/Thor 1200w PSU/Dark hero x570/new NVME drives. Started with the ROG Strix 6900xt LC OC, crashes start happening immediately when gaming with stable stress tests with absolutely never having any heat issues (CPU/GPU hotspot and base temps). I did take note that when the crash first occurs (random when the first one happens once gaming) that the subsequent crashes occur much quicker after hard resetting PC. Undervolted and underclocked the card with no improvement to crashes. I replace the new build with the 6800xt and 5700 xt, no issues what so ever. I'm at an impasse, because at this point the GPU has to be the issue and I know I bought it used...but I would like to try everything before I attempt an RMA, or will flat out pay to have it "repaired" because at this point I've invested to much to just let it sit.
Q1. Does anyone have a video/images of a breakdown of the 6900xt? I've seen some videos on YouTube but its just before and after breakdown images.
Q2. Would replacing the thermal pads/repasting the GPU even though heat doesn't appear to be an issue be worth trying?
Q3. I'm aware I'm not the original owner, I likely wouldn't get an RMA passed through but would be totally okay having it repaired. I do have the original invoice and tax invoice, they just obviously don't have my name on them. How is the process for just having something repaired? Are they timely? Should I just forgo even touching the card and immediately send it in to ASUS?
Q4. The one thing I've never monitored while stress testing and playing games is VRAM/other GPU components...could thermal pads mispositioned cause crashes like this to happen?
TL:DR Has to be issues with the ROG Strix 6900 XT LC OC, where to go from here as far as breakdown, replacing the thermal pads, repasting the GPU, etc.