Specifically a lot of people using Zotec and many other brands having to underclock their 3080s to avoid things like black screen issues and constant crash to desktop in games.  Has anybody that's managed to get any of the EVGA 3080s run into any issues like that, or did EVGA manage to avoid that problem?
 
Seem to see a lot of complaints elsewhere but not here.

da3wilson
Very good question. Wonder if it related to the specific brands or just a 3080 thing.

I'm not sure, but it seemed like a lot of people with a ton of different brands were complaining about it, but nobody with an EVGA card that I could see so far.  The consensus seems to be pointing at nvidia driver issue, but nobody really knows. But underclocking by 50Mhz seems to solve it.

- There were issues with the first few batches of 2080 Ti's.
 
- Not all cards are as good as the FE cards and who knows what the AIB's are sending reviewers, they could be cherry picked and their production cards are a bit soft in any number of ways.
 
- Lotta noobs out there with poorly setup and possibly new systems, with undergunned supporting parts (questionable psu, bad case airflow, etc) that can't just plug in a 300W+ card and run trouble free.  They aren't using them on proven, open air test bench setups and monitoring the card.
 
- There's a ton of benches out there on 3DMark that are on the bottom end of reviewer's results or even lower, look for yourself.
 
- Drivers, bios and chip quality are still getting refined.

Urieldagda

da3wilson
Very good question. Wonder if it related to the specific brands or just a 3080 thing.

I'm not sure, but it seemed like a lot of people with a ton of different brands were complaining about it, but nobody with an EVGA card that I could see so far.  The consensus seems to be pointing at nvidia driver issue, but nobody really knows. But underclocking by 50Mhz seems to solve it.

Interesting. The consensus I've seen is that no one knows what's really causing, and the best guesses are actually hardware faults with how the cards are designed.
 
The reason for this is that the issue persists even on 1200 watt power supplies, so it can't be a power issue, and these are non-oc at stock out of box speeds. It seems to happen most consistently when cards boost THEMSELVES above 2000mhz, and slight underclocking seems to fix the issue - this is why people doubted it was a driver issue.
 
Now, a new driver (beta? I don't know) seems to have fixed the issue, but many people simply think it's a hardware limit that the new drivers are now better at avoiding rather than a specific issue with drivers themselves.
 
It's interesting, at least. It all feels very rushed. Nvidia could have easily just sat back and waited for that mythical october stock that's supposed to fix all our problems, banish scalpers to the shadow realm, bring peace and love to the world, end global crisis, etc, etc, and then released in the middle of october or whatever.
 
Instead, they shrugged, decided giving some of the biggest retailers in the biggest demand nations like the U.S. 10 cards, threw gasoline on the tech tuber fire, and decided it'll just work out. 
 
Nvidia - a very old, experienced producer of tech - either has no idea what they're with supply, they have no sense of their consumers, or they are REALLY afraid of Navi and know something the rest of the world doesn't. Enough, I guess, to completely botch their launch and threaten the goodwill they obviously wanted to earn back.

Zotac has been the internets problem child as of late. PNY might be next in line for that position though.

HawkOculus
See the other threads related to the capacitors on the backside of the GPU? These clock issues seem related to the caps that are used on different boards.

Yeah I just saw that, though I'm not sure it's been fully confirmed or if anything can be done about it.

gravedigger78

Cerulean_Shaman

Urieldagda

da3wilson
Very good question. Wonder if it related to the specific brands or just a 3080 thing.

I'm not sure, but it seemed like a lot of people with a ton of different brands were complaining about it, but nobody with an EVGA card that I could see so far.  The consensus seems to be pointing at nvidia driver issue, but nobody really knows. But underclocking by 50Mhz seems to solve it.

Interesting. The consensus I've seen is that no one knows what's really causing, and the best guesses are actually hardware faults with how the cards are designed.
 
The reason for this is that the issue persists even on 1200 watt power supplies, so it can't be a power issue, and these are non-oc at stock out of box speeds. It seems to happen most consistently when cards boost THEMSELVES above 2000mhz, and slight underclocking seems to fix the issue - this is why people doubted it was a driver issue.
 
Now, a new driver (beta? I don't know) seems to have fixed the issue, but many people simply think it's a hardware limit that the new drivers are now better at avoiding rather than a specific issue with drivers themselves.
 
It's interesting, at least. It all feels very rushed. Nvidia could have easily just sat back and waited for that mythical october stock that's supposed to fix all our problems, banish scalpers to the shadow realm, bring peace and love to the world, end global crisis, etc, etc, and then released in the middle of october or whatever.
 
Instead, they shrugged, decided giving some of the biggest retailers in the biggest demand nations like the U.S. 10 cards, threw gasoline on the tech tuber fire, and decided it'll just work out. 
 
Nvidia - a very old, experienced producer of tech - either has no idea what they're with supply, they have no sense of their consumers, or they are REALLY afraid of Navi and know something the rest of the world doesn't. Enough, I guess, to completely botch their launch and threaten the goodwill they obviously wanted to earn back.

Botch their launch? I agree everything would have been smoother if they just waited to build a stockpile and release later. Not only are they selling every single one, they are the talk of the town. They released a 3090 to compete with themselves to make their own 3080 look like a bargain. 

They're being ridiculed for their 3090 and the it's not worth bragging if selling "every single one" means 10 cards a week, hah. Though definitely misdirected, the frustration people are feeling will turn to negative energy for Nvidia's brand, which is exactly why they apologized twice already instead of shrugging and deciding people are going to buy the card anyway. While your opinion of them might differ, Nvidia cares enough about tech tubers to make sure they get cards and even hooked up one of them with a $30,000 tv, and yet most of them at the most optimistic end of the spectrum agree Nvidia could have done a lot better.
 
It's obviously subjective, but by and large most people, from journalists to tech tubers and consumers, call this a botched launch. And it's annoying when people are quick to point out this isn't a paper launch, though most people are using it figuratively right now anyway. Giving one of the biggest computer vendors in the U.S. 10 cards while the largest online vendor still has people wondering if they ever had stock at any point yet, it's laughably not a paper launch only on... paper. That's exactly why they use ambiguous terms like "but we had stock" or "it was the same stock as before" while very carefully avoiding numbers or supply estimates which is effectively useless and tells us nothing.
We definitely disagree here, though the real kicker is going to be when next-gen rolls around and we get to see if Nvidia learned anything.

The 3080 FTW3 has 6 POSCaps, which is what they are saying causes the issue........

CandiceJoy
The 3080 FTW3 has 6 POSCaps, which is what they are saying causes the issue........

I thought it was the other way around and it specifically didn't have the bad capacitors. Didn't Jacob Freeman tweet about this?

Looks like the EVGA cards are all in the clear.

This seem to be a real issue:
EVGA: How are you handelng this for you´re hydr coppers and hybrid coolings? Seems kinda worthless investing in these if they will hade the same problem 

Poss273

CandiceJoy
The 3080 FTW3 has 6 POSCaps, which is what they are saying causes the issue........

I thought it was the other way around and it specifically didn't have the bad capacitors. Didn't Jacob Freeman tweet about this?

This is why they delayed the cards. They updated the pictures. It now uses the 4 + 2 method, though supposedly the problem is now considered widespread and that many people are having this issue, even those who were supposedly "safe."
 
It's being assumed that it might be more than just the capacitors, but I guess time will tell.