Unusual high failure rates for GeForce RTX 2080 Ti?

From the current posts out of 11 2080ti card referenced by user posts here, 1 has failed which is 9% failure for this thread. 4 of the 2080ti cards were the Nvidia FE cards of which one was the failure so 25% failure so far for the FE edition 2080ti cards and none for other manufacturers..., thus far.

 

I saw 2 users recently in 2080ti thread or was it 3, not so innocent if you ask me.

All were FE cards though.

Could very well be shitty vrm or something, I know custom use better ones, guess it's still so. Correct me if I'm wrong.

 

iirc an early board teardown at Gamer's Nexus indicated that the VRM were actually over-spec'd implying that power shouldn't be an issue. (this is from memory, I may be wrong)

2080ti FE here and no issues so far, reasonable over-clocker up to 2100/1000 (not sustainable long term though )
I have also tried the M$ DirectX RTX demos and no coil whine that some users over at reddit reported (fingers crossed I've got a good'un! )

 

2080Ti FE owner here, no problems to report so far. I had been waiting on water blocks to get here before completing a new threadripper system with the 2080Ti. I opted to drop the card (temporarily) into my old X58 system to test in light of problems people mentioned on the nvidia forums. With the stock cooler, I applied a mild OC (+100/200 core/mem) and still haven't seen any problems. FYI, 4k 60fps is very doable on X58 despite whatever cpu bottleneck is likely happening - that is both awesome and sad (for how far we have come in about 10 years).

Here's hoping the 2080 Ti doesn't start showing problems after I go to the trouble of mounting the block and adding it to my loop in a few days.

 

I'm fairly sure 99% of 2080 Ti buyers were buying a 2080 Ti for the first time. The remaining 1% would be those who later got another one to go for SLI, instead of immediately buying more than one. Otherwise you have no idea of their backgrounds. Why would they visit the Nvidia forums in the first place before they encountered a problem? They could have had 10 Nvidia cards before but if they had had no problems, their problems were with a board partner, or their problems were more straightforward, they would have had no reason to register or visit the Nvidia forums. Your judgement is too hasty.

I've had more Nvidia cards than Ati/AMD cards, but I don't have an Nvidia forums account. I don't know from which one my next card will be, but if it's from Nvidia, I still won't have an Nvidia forums account if I can help it.

 

Many people sign up to a forum for the first time when experiencing an issue - I did it on the ASRock forums a while back when I had problems with one of their boards. This is nothing unusual.

 

...and the subsequent denials about the denials !

 

Well, do you think everyone around has AMD/nVidia/intel accounts? I have none of those. For all I care, one can create account 15 years ago, be in comatose state and wake up.
Or one can be rocket scientist who never registered on any forum, nor posted anonymously single post publicly. But I am pretty sure that when you pay that kind of money for GPU, you'll more likely to go and ask manufacturer in case of troubles before asking for RMA. Which may require to create new account for great many.

 

Is it possible that these problems are being caused in the long run due to the 'professionalization' of overclocking? Earlier, manufacturers used to caution their customers about touching anything, and overclocking was a thing of select few. The safety margins were larger.

Now, it is as if, he who doesn't overclock shall be smited of the face of the earth. Everyone is taking these risks now, some in reasonable amount, and some going overboard.

Consider the bullet trains (or even other vehicles), they have a higher theoretical speeds, but are run at around 75-85% speed at max, and that too for only short durations/stretches.

 

My EVGA RTX 2080 Ti XC Ultra working just fine .. thankfully

 

found info about what batches/aka S/N are good, so...good are batch 0333xxx. If u have 0332xxx then probably u fked up. Unless u have galax card, which has loosed mem timings, so probably can last longer...

 

RTX 2080TI FE owner since Oct 8th. No issues as of yet except some drivers work better than others but Nvidia also did hotfix driver for the same driver version that was causing crashing in Destiny 2 for me. I believe 416.16 was the only that really caused any random reboots or crashes for me. My setup consist of a Gigabyte Arous Gaming 9 with a Intel 7820x @ 4.7ghz and 32gig of 3600 gskill memory. The card overclocks 158 on core and about 500 or so on mem and I get just over 14000 on 3dmark.

 

I have a Gigabyte Gaming OC 2080ti, have had it for about 2.5 weeks now with no issues so far. I haven't done any additional OC'ing since I opened the card out of the box, I only use it about an hour a day some days not at all.

 

I purchased two 2080 Ti FE cards and an NVLink bridge from NVidia.com. Prior to installing these cards I had two 1080 Tis connected with an HB Bridge. I installed the 2080 Tis, taking all the normal precautions (been doing this for many, many years without issue). With zero overclock beyond the way the cards were configured upon receipt, I ran FireStrike Ultra and Extreme, VRMark Blue Room (131 fps, way beyond the necessary 109 fps required to pass the test), I was getting 133 fps in Shadow of the Tomb Raider at 4K highest. As an aside, the ambient room temperature is 68 F. I was exhilarated at the performance of these cards. After benchmarking for about an hour, I exited all running programs and my system was sitting at the desktop. I left the room for maybe 30 minutes - when I came back, my system had rebooted and I my login screen was huge, 1024X768 or maybe even 800X600. I logged into Windows and when I right-clicked desktop and selected "NVIDIA Control Panel" I was informed that my display was not connected to an NVidia device. Now, normally you'd see this if you connected your display to integrated graphics instead of your discrete graphics solution, but this wasn't the case - my display port cable went from my 4K 120 Hz G-Sync monitor directly into the top-most 2080 Ti. I tried various things (doing into Device Manager and uninstalling the top-most card and then scanning for hardware changes (Device Manager showed a warning icon next to the card description but the 2nd card didn't have the warning icon). Anything I tried, including booting into Safe mode, would result in screen corruption and an eventual BSOD. I removed the top-most card and left the 2nd 2080 Ti in it's slot. This card seemed fine at first, so I started the RMA process on the first card. The next night, after reading about reported 2080 Ti failures form other buyers, I decided to run a few games on the remaining card. I started up Doom (2015) and at 4K with everything maximized, I played for maybe 20 minutes before my screen started widely artifacting - and then Windows froze and I had to reboot.

Both cards failed in a matter of a two days with the same symptoms.

I have since requested a refund and my cards are on their way back to NVIDIA as I write this.

This, for me, was hard to do, as while I had the cards in NVLink, while they were working, it was the most amazing experience I've had in all my years of PC gaming, with regard to blisteringly fast frame rates.

When the first card failed, I resigned myself to the RMA process and decided I could "live" with one 2080 Ti while I waited for a new one to arrive - after all, most recent games don't benefit from NVLink (SLI) - but when they do, the scaling can be almost unbelievable.

 

I was looking for help on this issue for a while, and I am glad to see that I am not the only one with this issue. I pre-ordered a 2080Ti FE on day one. (Direct upgrade from a 1080Ti) The card arrived and worked fine for a few days. Eventually, the card failed to wake from sleep and randomly would not display an image when doing a cold boot (black screen).

After that, I started to get weird dots on the screen in Deus Ex Human Revolution. Then games started to randomly crash to desktop.

I started the support chat process with Nvidia, who wanted me to run diagnostic software. I said screw it and just decided to return it. Since I decided to return it and not play the RMA game, I was required to go buy my own return shipping and insurance. So, now I'm actually out money on this deal. I also sold my old card, so now I actually don't even have a video card. I'm not sure what to do now, but I will probably try buying a different 2080Ti like an EVGA when they are in stock.

 

There are "special sauce" drivers that enables you to still game with quad sli... but it's kept quite hush hush.

 

Well as always, sli is pcie bandwidth and cpu limited - quad sli even more so. But in frostbite engine games, performance is excellent with 40+ pcie lane cpu's.

And yeah, it needs to be applied with each new driver.

Quad sli died with 1080 ti's though, as you can't physically use more than 2 gpu's with nvlink on the consumer gpu's.

 

I wonder if i will be happy when i will finally switch on my 2700x vs i7 990x that i m replacing. I could have bought just the 1080ti to replace the old 2 x 770 gtx maybe

 

What diagnostic software Nvidia wanted you to run?

 

For me all they asked for is for me to run msinfo and send them the export