RTX 2080 Ti Failing Too Early

[AlfaProto]

Source is coming from:

It's difficult to get a quote on which part of the RTX 2080 Ti died, however, users claim to have failing RTX 2080 Ti, even running at stock

Quote

Mine had faulty memory, it died within a month, hadn't even OCed either.

This affects AIB 2080Ti too:

Quote

Not just the FE cards. Zotac 2080 ti's are having issues as well.

However, it should be noted that users reporting RMA are mostly the RTX 2080 Ti:

Quote

Most of the reports seem to be from 2080Ti’s so it’s possible I just got unlucky

 

In my opinion, I think this is just another luck with the bath tub curve model. Early model/revision are quite buggy, defective, and/or prone to fault. I would like to add that this is the first time where a x80 Ti was released alongside with the x80 and x70, instead of after a Titan series. 

 

What do you guys think? This is an unlucky struck or Nvidia just rushing their cards out into the market, despite game developers are working on implementing DXR into their games?

 

Please, I would appreciate that you remain civilised in commenting this news.

 

UPDATE on October 31, 2018

DigitalTrend brings attention to the RTX 2080 Ti dying prematurely. Source.

Quote

Threads have been cropping up on Nvidia’s forum about dead and dying RTX 2080 Ti cards for weeks now, with almost every thread filled with hundreds of comments highlighting crashes, black screens, blue screen of death issues, artifacts, and cards that fail to work entirely.

While the RTX 2080 is also affected, it seems that there are more reports on the RTX 2080 Ti than the 2080.

Quote

There are some reports of issues with the 2080, too, but the majority reference problematic 2080 Ti cards.

People are reporting this issue on Reddit threads.

Quote

Reddit threads with a similar theme have been appearing as well, detailing the RMA process that many users are now going through.

 

[mr moose]

I had a quick read through the reddit thread.  Are we talking an handful of people all with the same problem or is this likely just the usual failure rate?

[AlfaProto]

1 minute ago, mr moose said:

I had a quick read through the reddit thread.  Are we talking an handful of people all with the same problem or is this likely just the usual failure rate?

It's difficult to point whether this is a failure rate, or Nvidia rushed the model into the market. Not really, there are other users reporting 2080Ti failure at other forums, and RMA'ing it.

[D13H4RD]

It's probably within the margin of defects when it comes to mass produced items, to be quite frank. 

 

AIBs and NVIDIA should be replacing them through RMAs as per the usual. 

 

If there is actually a larger than usual number of defective units out there, then we have an issue 

[Nystemy]

Defects are expected, everything can fail after all.

And since it is a new chip in the sense of things, then Nvidia might still not have their quality control fully tuned yet and some defects aren't properly identified. This is normal for new releases, especially if they are a little rushed.

But Nvidia has most probably accounted for this and are most likely eager to fulfill RMAs for the one's suffering from chip yield.

Also, memory issues can be expected if the chip talking with the memory isn't working correctly. It is more likely that it is the GPU chip that has poor yield during early manufacturing, rather then a RAM chip that has likely been on the market for many months.

[GoldenLag]

2 minutes ago, VegetableStu said:

oh damn, if it's the VRAM it's not good news for micron (if i remember right) o_o

It shure as heck will have some Ram-ifications

[Taf the Ghost]

Likely just more vocal user base for a high-end GPU like the 2080 Ti, so it should be expected we'd hear about the failures more often. Though it is a new model on a brand new GDDR memory with a lot of it clocked highly. It's very possible this could have issues beyond the normal ones. Especially if some early memory batches might not have been binned properly.

[mr moose]

12 minutes ago, AlfaProto said:

It's difficult to point whether this is a failure rate, or Nvidia rushed the model into the market. Not really, there are other users reporting 2080Ti failure at other forums, and RMA'ing it.

So at the moment its just normal failure rates then.

[schwellmo92]

21 minutes ago, VegetableStu said:

oh damn, if it's the VRAM it's not good news for micron (if i remember right) o_o

The guy has nothing to qualify his statement but if it is true I agree it’s bad, first mass production GDDR6, not even hitting the original projected speed (16Gbps) and then having defects to top it off.

[mr moose]

2 minutes ago, D13H4RD2L1V3 said:

Pointing at the door behind her 

Peek and poke are the basic instructions accessing memory.  I was trying to be funny,  sorry.  I'll see myself out.

[leadeater]

4 minutes ago, mr moose said:

Peek and poke are the basic instructions accessing memory.  I was trying to be funny,  sorry.  I'll see myself out.

[D13H4RD]

Anyway, back on topic. 

 

Definitely sucks when your $1000+ GPU is failing quite early, but I'd assume that AIBs and even NVIDIA should feel the same. 

 

At the very least, they should be replacing them without question. Perhaps later on, the failure rate should be lower. 

[Taf the Ghost]

It a free for all on RAM puns, or more of a random access thread?

[leadeater]

Since it's sounding like memory errors then I'm not surprised being that GDDR6 is very new, would of expected 2080's as well though. Maybe it's memory controller faults rather than the memory itself.

[asus killer]

so the cards died without even raytracing. A life without raytracing, a life without meaning. That's just sad  

[Delicieuxz]

There were also some other reports of bad 2080 / 2080 Tis posted on HOCP forum:

 

https://hardforum.com/threads/the-nvidia-rtx-2080-experience-signs-rtx-off.1970174/

 

Quote

Yeah. I ordered two. One card wouldn't even get to the post screen. The other card worked, but the LED logo light wouldn't light. I RMA'd both. Yeah, the second card worked, but for $1200+, I want a fully working card, which I don't think is too much to ask for. The third card arrived, and I was getting artifacts.

 

You can see a picture of both cards. (I kept the working card with the broken LED until I had two working cards).

 

 

SLI when taking this video, but after testing only the first card, that worked, and the second card still had the artifacts. And I couldn't start any games, which would freeze within seconds. I'm now on cards 4 & 5, which are working, and hopefully will continue to work.

 

And:

[MMKing]

People with malfunctioning hardware are more likely to post about it. Still, these are not good early news.

 

33 minutes ago, leadeater said:

Since it's sounding like memory errors then I'm not surprised being that GDDR6 is very new, would of expected 2080's as well though. Maybe it's memory controller faults rather than the memory itself.

Could also be the 2080Ti has a higher failure rate on the account of having 3 additional memory chips.

[laminutederire]

1 hour ago, leadeater said:

Since it's sounding like memory errors then I'm not surprised being that GDDR6 is very new, would of expected 2080's as well though. Maybe it's memory controller faults rather than the memory itself.

No one buys the 2080s so we may never know.

 

As we can see it in this thread, at this point Nvidia could fuck up, people would find an excuse for them as if everything is normal. Too much blind faith in them.

I'd be interested in how much rma'ing those cards will affect the already behind manufacturing schedule with demand over the roof compared to the supply. Or will they bin cards less so that they can push more of them to demand now? Tilme will tell

[mr moose]

8 minutes ago, laminutederire said:

As we can see it in this thread, at this point Nvidia could fuck up, people would find an excuse for them as if everything is normal. Too much blind faith in them.

That doesn't even make any sense. what blind faith or excuses? for what even? 

 

[leadeater]

6 minutes ago, mr moose said:

That doesn't even make any sense. what blind faith or excuses? for what even? 

You know how it goes, talk about something from a technical aspect commenting on how it works or why it might be happening and that is no different to being biased/shill/blind faith/defending said company etc 

[mr moose]

2 minutes ago, leadeater said:

You know how it goes, talk about something from a technical aspect commenting on how it works or why it might be happening and that is no different to being biased/shill/blind faith/defending said company etc 

Did I commit the crime of not trying to jump to conclusions?  

[laminutederire]

1 hour ago, mr moose said:

That doesn't even make any sense. what blind faith or excuses? for what even? 

 

Everyone says here nah it's normal failure rates and so on, when the only defence valid at this point is to wait on objective numbers to know the actual scale of it, instead of just noticing that there are more posts and faster compared to the 1080tis. But no people seem to think it's better to say it's true but it's normal failure rate and so on.

By the way my answer had nothing to do with what @leadeater was talking about. I was answering to him about why we don't see much about 2080s which is to me because it's not being bought enough for it to be noticed, unlike the 2080ti which actually has a success since it makes more sense.

[leadeater]

20 minutes ago, laminutederire said:

By the way my answer had nothing to do with what @leadeater was talking about.

Don't worry what I said had nothing really to do with what you said either, @mr moose knows who I was hinting at though 

[Carclis]

3 hours ago, asus killer said:

so the cards died without even raytracing. A life without raytracing, a life without meaning. That's just sad  

At least RMA-tracing is working

[D13H4RD]

5 hours ago, laminutederire said:

As we can see it in this thread, at this point Nvidia could fuck up, people would find an excuse for them as if everything is normal. Too much blind faith in them.

Uh.......huh.

 

I reckon that if the same thing occurred with AMD cards, the response might be the same. 

 

Some people aren't so quick to jump to conclusions.