Hello everyone,
I just bought a brand new RTX 2080ti FTW3 only 5 days ago (I was happy for my first EVGA card), but it's already dead :(
I thought it was safe now to buy an RTX card after the launch 6 months ago, but what a mistake!
Everything was working fine until today when my PC suddenly crashed with a hard reboot (all temperatures were fine before the crash).
After trying to boot several times, Windows could finally start, but at the very low screen resolution, and saying in the device manager that my card 2080ti is not used because it has a problem:
Since then, my PC reboots randomly with artifacts:
Now I have a 1.500€ RTX garbage in my PC :(
Why do we have to endure this nightmare for this RTX generation with no explanation from Nvidia?
With all the hassle with RMA process, and the risk my package could be lost during the delivery, or receive another defective card.
I am so upset against Nvidia by their disregard for their customers...
Edit: add more information
I have tried to revive this "Frankenstein" RTX card, and in the Windows device manager, when I try to activate it, I get instant crash with blue screen, core dump and an error message about Nvidia module "
nvlddmkm.sys":
From my own experience, it's like the chip itself has collapsed suddenly!
Just out of curiosity, at the beginning (before the fatal crash) I have tested the Micron memory and it looked pretty strong, even able to run at +1000 mghz (16gbps instead of 14gbps).
My guess is if the problem was only the memory, then my card would still be able to run some how a little, even with errors or corruptions and artifacts (like with defective PC RAM)
It's hard to believe that 11 modules of memory could vanished at once!
So I guess it's TSMC contamination which is causing all this troubles, explaining why each card has a different behavior and different surviving time because of different contamination level! And why cards with Samsung memory are less prone to failure, simply because they have been integrated later.
Nvidia should have recall all this defective batch, instead they chose to let the customers clean all their bad inventories with RMA until the last one...
By the way, my S/N starts with 18, which means made in 2018.
I guess there are still stock from 2018 on the shelves from around the world (more or less contaminated by TSMC) waiting to die (mine bought 8 March 2019)
Edit: it's even worse than I thought because I have found on Nvidia forum a defective card from 2019 with Samsung memory:
Can't insert link, but the topic is "
brand new evga 2080ti ftw3 iCX2 hydro copper... error 43" (with a GPU-Z screenshot puu.sh/D0WTA/3383729c32.png)
I believe Samsung and Micron are not beginners to make both defective modules GDDR6.
Which tends to confirm that the problem is the core, and probably related to the contamination at TSMC fab 12nm.
So we know that the memory brand doesn't matter, and the problem reside in the core itself.
And we know that even cards from this year 2019 can die too, which is not reassuring, especially without any clear statement from Nvidia...