Unusual high failure rates for GeForce RTX 2080 Ti?

I believe you bro.
But if you want to be taken _somehwat seriously, you need to lose _something. Either troll face, AMDfan nick, or attacking legit owners in the thread about the card you have no interest in.
I mean taken all at once, it's a little bit too much. Almost full r3tard

 

We don't need some random on the internet to warn us about anything.

Most of us have been around long enough to make our own informed decisions about things like this and don't need 'advice' from someone who likely has zero experience with these cards.

Thanks for caring all the same

 

in a computerbase topic i found an interesting post, it compare the distance between the VRM area and the Vram in a FE PCB and an MSI GAMING TRIO X PCB
https://www.computerbase.de/forum/t...rce-rtx-2080-ti.1832694/page-20#post-21887316

 

Good suggestion. As I wanted to keep these cards quite desperately, I tried a number of things I didn’t mention in my original post. First, I took the first card to fail and put it in another one of my gaming machines - one with a single-rail 1200 Watt power supply (about 8 months old). It exhibited the same issues. When the 2nd card failed, I originally left it in the 2nd slot, but in the middle of the night, able to sleep due to these issues, I put the 2nd card into the first slot, i unwrapped a brand new 1600 watt EVGA Supernova power supply that I got on sale at Microcenter a week or so ago (just in case my current power supply wasn’t powerful enough) and reconnected everything. Same issue. 2nd card failed in the 2nd pc with the 1200 watt PSU as well.

 

I am not saying that anyone is lying but all these complaints seem fishy to me. There's been quite a number of people making it a point to bash Nvidia and Intel. How do we know that people weren't overclocking their cards? I think that when people make complaints like this in public they should take a picture prove that they own but they're complaining about. That would at least make the complaint seem a little more legit. I also have a 2080 Ti and after the complaints I started running heavy calculations on mine and it's been running now for a few days non-stop. The GPU is running at 75% usage and the temperatures have been a steady 60 or 70 degrees Celsius and the RAM usage has been about 6 to 7 GB. That is running its stock but I did use the software to put it in gaming mood for a day which didn't interrupt the calculations. So far my computer hasn't crashed and it seems like the calculations have been uninterrupted based on the amount of tasks being done. I know that that's not a benchmark and maybe because I'm not using something that's graphical like a game the results may be different than other people. I bought the Asus 2080ti OC Edition which is white and black with the black and white back plate. I'm using the latest drivers as well and also I reinstalled Windows on this particular computer. I'm not at home right now but if I can remember I'll edit this post and take a picture of my card if I can because I really don't want to take it out of the computer right now because I want to finish my calculations. anyway I don't really think this is something people see get worked up about even though it is inconvenient for some. Whenever I buy something this expensive I make sure that I get an extended warranty or extra warranty so that I don't have to worry about sending it through the mail. I just wanted to Microcenter and they give me a new card or just give me my money back so I can buy a new card. overall the main reason why I take these complaints for the grain of salt is because so many complaints about the RTX release so no real good reason.

 

There is the 'echo effect'. Similar to 'all EVGA 1080s were catching fire' when it ended up being a couple of videos and a bad batch of VRM components that failed on some. It gets picked and makes the rounds of the tech media and 5 examples become thousands. On the other hand, this is a complicated device and their first effort at this design pardigm so I would not be surprised to see a somewhat higher failure rate than from a more mature process. We don't have the numbers to know.

 

That could be, but when actual system builder says that 3 out of 9 cards they received were faulty and some reputable reviewers says that they got also failed cards, it does not really looks like its just few random cases.

And statement NVIDIA put says nothing, just regular nothing saying pr statement. When NVIDIA says few... it might be many thousands and they would still say its just small amount (out of all cards they sold per year).

 

I have the OPPOSITE experience, coming of AORUS 1080ti with 2.7/2.5 heatsink to gaming oc 2080ti with 2.0 heat sink and i get better temps!
Played some wolf2 with all settings maxed and 120fps limit to see how its 4k game with so high fps, and never passed 60, when i had vsync 60 it was running in 50ish celcius

 

Well you could just check 2080ti owner thread, but no, make theories...
I counted at least 5 guys here with defect/dead cards.
I know another enthusiast from my country, who's both card in nvlink died, one after another. Other guy starting artefacting(memory downclock helps, so it is memory problem)....so....

 

Nothing works this way. Vega cards should be DOA, same for Fermi, or any console.

Turing has a ton more stuff going (mainly hardware scheduling features), part from Rtx and the tensor cores.

 

I have zero plans when it comes to owning one of these cards. That being said. Nvidia will have to fix/replace your card and Nvidia will have to correct whatever is causing this issue. If they do not, they will not sell any 2080ti's. it's pretty simple.

 

hear! hear!

EDIT: It's good the folks who aren't having problems are posting as well please don't slam them for it; They're helping in the troubleshooting / isolation process of establishing a pattern against defective cards. I would just add if possible also post full specs - (Make/Model - batch #) I'm trying to articulate something along... collecting a sample size of data comparing to defective gpus... which NV will be looking at... to inspect, hopefully. Since it's obviously not all of them/per all PCBs per all dates etc etc..

 

You do realize the maximum temperature for a GPU is typically around 105c? Sure, you don't want to be getting close to there, but 80-90c is well within operating temperature.

 

95c is also the peak operating temperature for microns GDDR6 chips

 

Wanna see something by this guy.

 

EDIT: M6 and M7 modules of the GDDR6 failing (cannot sustain ~96°C at 100% load)

by pcgamesn.com

 

They've started an Owners ROH thread for RMA'd 20 series cards on Overclockers now

 

https://www.hardocp.com/article/2018/11/09/geforce_rtx_2080_ti_fails_after_gaming_for_2_hours/

 

so after watching that video, i can tell you the issue is not the memory chips themselves.

something is going wrong in the gpu's imc or cache, these are just your typical large die failures.