Looks like the poll wasn't too far off after all...
http://www.nvnews.net/vbulletin/showthread.php?t=114925&page=17

The stats so far:
http://img232.imageshack.us/img232/9134/failure814ed3.jpg

My desire to make my next update a green one is dying by the minute!

25% failed acording to this poll
assuming, that multiple = 2
I get 25% failed shipments.

I have to say that I had 100% (5/5) Asus failures, thus I would care for a poll in that direction.

Put an Accelero S2 cooler and a slow 80mm fan on my 8800GT almost immediately, and it's at around 55C when working. No problems yet.

I have to say that I'm a bit puzzled by something. Every time AMD reports a loss, TR is usually right on the story splashing it with a headline. But yesterday, nVIdia reported a $120M quarterly loss, and a drop in revenue, and so far I haven't seen anything mentioned about that at TR. Well, it was only yesterday, and this may be nothing but a timing issue for the site's personnel. In looking at a number of other sites, though, who also seem quick to headline AMD losses, I also can't find the story. Here's the link:

http://www.marketwatch.com/news/story/nvidia-reports-huge-loss-stun...tory.aspx?guid={E137EFF1-8220-40EC-B324-366D9957...[watch.com]

Or, you can just Google "nVidia quarterly loss" and see where it shows up.

I know I had a 8800GTS 512 fail on me. But mostly due to the driver not spinning the fan up when under load. eVGA replaced it and with lots of fussing and riva tuner I was able to get the fan to move air correctly and it has been fine ever since.

Still that should not be a problem out of the box.

Perhaps a run through with Driver Cleaner?
Reinstall with a previous set of known good drivers for your card?

does anyone here know how to put a permanent end to my BSOD nv4_disp.dll error in my 9600gso? its horrible and other than a single 5 hour gaming session , i have not been able to play anything for over a week despite update to 177.83??!!!!

Inquirer is known for its low standards, even before this nVIDIA thing; they publish multiple articles, first speculating, then contradicting it, and every possible outcome, so they can say, we told ya, you read it here first.

it used to be more Intel vs AMD, but there's not much there now. last I read, they claimed that AMD [Phenom] will end up faster than Intel, I don't follow it anymore...

I've had a few different NVIDIA cards fail on me at work, but I use so many cards its hard to say whether it is anything unexpected (say 2 or 3 cards out of the 30 or 40 I've used over the past 6 months or so)

My 8800GT hasn't failed on me, yet, but I won't buy another Nvidia product again for a very long time. Their driver team is completely useless.

Not only is their control panel so bloated and horrible I finding myself even missing the good old CCC. But I haven't had a single driver that didn't have some really stupid bugs in different places for a good while now. And their installation puts several annoying context menues (control panel, play on my TV, etc) in place aswell, of which some can't be disabled without going into the registry.

Currently on my third Nvidia video card on my computer in a year. Ati is looking good for the first time for me since that crappy 8500 I had.

My (G80) 8800 GTS 640 is sweet. But I'm thinking my next could definitely be a 4870.

Does anyone know what the exact problem is?

Someone from ATi told me that because the substrate (PCB) uses is RoHS, but the solder balls on the chip still have high lead content, constant thermal cycles causes them to merge and crack over time.

Scott - An idea: Next time you post a potentially inflammatory poll, please keep the results closed until the poll is. I don't think anyone here would, but showing the results after each vote might encourage ballot-stuffing on a fanboy-sensitive issue like this one.

OMG..hundreds of failed NVidia cards in the poll already ....yikes!

My 9600 GT runs 24/7 ... never misses a beat. I'm happy with it.

It seems that the whole problem if any is probably linked to a certain batch of chips.

I want to see the dates on "defective" chips if there is any. There is where you can start to make some correlations.

I suspect this problem to be with TSMC and perhaps Nvidia's QA being too lenient on yields.

a cooling algorithm? i want to find the bit of math that makes my chips run cooler!

Please note that the first 8800GT (G92) BIOS release only ran the fan at a fixed 29% of maximum speed until the GPU core was > 90 degrees C. This was a ridiculous mistake, corrected by nVidia and (at least) eVGA and BFG about a month after shipment. The updated BIOS is download available from the manufacturer websites. With the updates BIOS, the fan speed rapidly increases above a GPU core temp of ~ 73 degrees C, in line with the fan-speed profile on previous nVidia GPU families. Remember that all commercial-grade silicon is standardized for long-term reliability at a maximum case temperature of 70degrees C --- about 80 degrees C core temp, typically. Regardless of whether it is ATi or nVidia. All of the ATi and nVidia silicon comes from TSMC and comply with TSMCs design and thermal dissipation rules.

Remember that the failure of the nVidia parts in the laptops was solely due to the failure of the bonding material between the chip and header, not due to the silicon itself. How widely that bonding material was used, only nVidia knows. From the information currently available, it seems that nVidia was aware of a potential problem with the substrate bonding well over a year ago, but collecting all the information to compute a $200million write-off in conjunction with the laptop OEMS on a time/temp -related failure would have taken a lot of time after the problem was first confirmed. Very likely that none of the latest 8xxx/9xxx families were still using that particular bonding materail.

Scott, it must have been the back-ache that clouded your judgement with regard to even considering taking anything from the (National) Inquirer at face value. Your informal survey is also likely to be contaminated by failures due to user-overclocking -- or in the case of the 8800GT (G92), the failure by users to update the erroneous BIOS in the first shipments.

Obviously, this poll is not scientific, and there is no way of knowing if someone is spamming it either.

Further, 12% failure rate is pretty high for a consumer product.

I had an 8600GTX fail on me after a week last year.

I have no idea what code name it was though.

"Charlie over at the Inquirer"

I stopped reading after that.

Personally I hope Charlie keeps sticking it to them until they come clean. In the mean time, I wish I had gotten the Intel X3100 graphics instead of the NVIDIA I got in my Latitude D630. At least then I wouldn't have to worry about this thing crapping out on me when I need it most.

My desktop 8600 GT (G84) is still going strong after a year of use.

If you take out that last option, and based on the current returns:17% report problems, 83% report no problems. All things considered, including the population being sampled here,I wonder if that represents a genuine trend.

I own a 7600GS and I am having troubles leaving my machine on overnight. No matter what I do, the video drivers fail each and every night - for the past 5 months now. The card works fine in my Linux machine, but not in my XP machine. It is REALLY annoying to have to restart my working machine every morning.

AMD GPUs = hot

Nvidia GPUs = fail

Is the last poll option really necessary?