All nVidia G84 and G86 cores are defective?!

No offense but if the OP and the author of that article are ATI stooges just trying to hurt to Nvidia because they don't have official information then what does that make the people that defend Nvidia with nothing but their own personal experience and biases. I've personally purchased 5 Nvidia gpus (even a FX series card when many considered them garbage) over the years and only 1 ATI card and I recommended a friend to buy a desktop 8600gt since he only wanted to play WOW so I'm definitely not biased towards ATI. But come on something is wrong and nobody knows what exactly or even which cards, because Nvidia won't tell us, so don't start with the fan boy stuff and posting even more mis-leading info.

 

No offense taken

Actually yes, the problem is well known: overheating. And it's mainly a cause of poor build quality of the host notebook/poor management by software.

 

Exactly. nVidia are not helping their reputation, they release a statement claiming they have problems with their last gen GPU's, then say nothing.
I'm not a NV "fan boy", but there is no point in believing the inquirer's report until it is either confirmed or denied by those responsible, your notebook manufacturer being the one we should be pointing our attention at.

 

Yeah I agree with you there. It's possible that cards heat more than they should because of a manufacturing problem but still, it shouldn't be the only cause.

What I find strange though: What were the OEM's engineers doing during the tests?

 

I would imagine they tested them with the older drivers. These did not stress the GPU any where near as much as the latest ones do. eg: 101.08: max temp 82C, 167.59: max temp 96C. These are the official drivers released by Zepto so I think they are representative. It's a big difference none the less, just a shame I want the performance the newer ones give .

Edit: these are my readings not anyone elses.

 

Based on reports that the problem is with the packaging materials, I don't think the problem was only extra heat. I think the problem was the cooling cycle, where you heat up the chip under load, and then turn off your laptop to move it somewhere else. If the packaging material is defective than these repeated cycles will eventually cause fractures and failures. If the voltage was increased as The Inquirer reports and the GPUs ended being hotter than anticipated, than this would likely exacerbate the problem. Presumably the OEM engineers found that the chips were hotter than anticipated but still generally within the maximum allowed temperatures. Since the G84 and G86 were released more than a year ago and the amount of failures is really starting to become obvious now, the engineers probably didn't place a year's equivalent loading and unloading on the GPUs to see if the failure rates were higher than normal.

And I posted the following on the Apple forum thread on possible G86 defects and I thought I might as well copy it here too. It's just my option on the possibility of a recall:

http://forum.notebookreview.com/showpost.php?p=3602346&postcount=21

 

No, all these complex theories are doing nothing but confusing people and making them think most of the 'article' is true... on what fact is the whole concept that nvdia have screwed up their packaging material based on? He just keeps jumping to conclusions based on stories he got from multiple sources saying completely different things? There's no evidence of any of that BS even being true!

there's just no way that all of the them are defective! they would've realised a LONG time ago... it's quite ridiculous that everyones trying to come up with theories based on top of this, it's only a matter of time before this thread spirals out of control becoming the source for more speculation and confusion...

nvidia announced that they shipped a batch of defectives... the defect being overheating. notebooks with good cooling might offset it right there... notebooks with poorer cooling systems in hotter than usual conditions might overheat and fail, being under warranty might mean the loss they were talkin about. It's probably a billion dollar buisness and for some jerk to sit on his desk and think he's come up with the answer to everything is just pathetic. there are just too many variables for anyone to speculate and post theories.

now, how the hell did that guy say all the gpus are defective and that the heat cycles cause damage in the long term? of course there are failures on these forums! but that's because no one comes and posts a thread saying 'my 8600m GT is working as expected, so i'm happy.'! there are definitely many many many more g84s and g86 gpus running fine!!!! it's just the nature of such forums that a majority of the reports are of failures and what to do about it, very few or none are about something working well.

Till anymore information is official, it's absolutely foolish to speculate that all the chips are defective. ridiculous!

 

You are aware that packaging material failure is official information aren't you? From an SEC filing.

http://biz.yahoo.com/e/080702/nvda8-k.html

Thermal and packaging problems have been officially announced by nVidia. The question is which specific products/batches are affected.

 

actually, I did read that... it was the first release right? I'm sorry i got carried away there, but what I meant was how they could screw up the material on all the chips... there's just no way. we'r talking about a company that's been making these products for a while and know the materials they'r using and they would've been put through their testing cycles... of course they released a statement and everyone knows that a considerable number of defective batches were shipped, unlikely we would've heard about this if it wasn't a big number. but suggesting that everything they've shipped to date is defective is stupid. there's just no reason to assume everythings defective till they announce what's actually going on... sure i could be wrong when they decide to talk next, or i could be right. but seriously, speculating that all of them are defective with absolutely no source of evidence is quite stupid!

 

Yes, I too find it hard to believe that the problem would be that large. I wish nVidia would have more to say about this, but I think they've been hit by too many issues at one time. The just launched GT200 series is a huge chip and they are likely still trying to ramp up yields while focusing on the upcoming 55nm shrink. At the same time they are still trying to launch new GPUs to fill out the 9M series. While ATI's new RV770 series is putting price and performance pressure on nVidia's entire desktop lineup and nVidia is trying to rush the 55nm shrink for the desktop 9800GTX+ into production to counter. And now they are hit with a large number of defects on their older mobile products. Their PR guys must be working overtime trying to calm things down.

 

There are like 300 people in this thread freaking out about their video cards, when not a single person here has had a big failure.

This is exactly the kind of crazy that's gonna drive us all mad.

Just wait for better drivers or fixes from companies that will allow the fan to kick in sooner.

And if something really is wrong, my Powerpro 3 year warranty is waiting.

 

for those who have defective chips and no extended warranty, not saying all are defective, forcing the fans to kick in sooner isn't going to help if you are gaming.

 

I totally agree with carthikv12.

There's no way all chips are defective, if all these rumors are true notebook manufacturers must have ceased production of all notebooks with g84 and g86 chips long time ago. Even HP,Dell and Sony are still selling products with these chips, it makes no sense to produce new defective products if the manufacturer will have to replace them with a new and more expensive one.

I own a VAIO FZ-250FE notebook with a 8400m GT GPU, it is 6 months old and I turn it on and off several times in a day and there's no sign of problem with it and as far as I know Sony have not released any BIOS update to solve any issue with the chip.

I point this to be a manufacturer related issue and given that all products are part of a mass production scheme just the way thermal paste is applied to the GPU and heatsink can be the reason of a +/- 20 [°C] difference in the GPU temperature in two identical notebooks and consequently a broken notebook in the short/long term.

It may be possible that first generation chips could be defective but you should stop trusting sensationalist reports/articles from no confident source.

Shoryu.

 

Ok, so some people may have bought a luxury item with a problem. After your warranty runs out, If it fails, you'll have to buy a new one. The same with all consumer goods from cars, to TV's and PC's. It's the way it is.

Stop complaining, and get over it.

 

Ill say that on my dell 1420 laptop, my geforce 8400m gs overheated from somehow getting some high temp from casual use just 1 month shy of my warranty expiration. After that, it was never the same and whenever the vid card temp reach 67 degrees celcius most from games or video card apps, the card would either crash or experience very degraded performance till it would eventually crash.

Sadly it took 2 dell depot shipments for them to actually do anything about the problem, since they're diagnostic software couldn't find anything wrong with the computer till the 2nd time where the card just died from normal use, which wouldve sucked if i ran out of my warranty.

Just got back the laptop, but probably the same old perhap defective vid card that may have temp problems.

 

Now that I think about it, this might explain why Nvidia released the slightly improved mobile 9 series cards. Maybe the whole reason for the early mobile 9 series cards was to fix and optimize the early 8 series cards which we are told to be defective.

 

exactly what i said before. it was weird that they would re-introduce the 8600gt as the 9500m. they must have knew that there were heat issues with the 8600gt

 

But when there's no one who knows that, why should nVidia admit the problem. Ruining theirself is the best to give ATi the lead.

 

Having the fans come in sooner may prevent some of these people from reaching crazy high temps whilest gaming.


And when I prop up my laptop it lowers my max temp by 4-6C and the average drops as well.

I feel as if I bought a Zalman 1000 or something it could do wonders..

 

I'm actually starting to regret ever starting this thread. The intent of the first post was just to point people towards a new piece of information. Considering that the initial statement from nvidia was posted on NotebookReview.com's front page I thought the article from TheInquirer.net would be of interest to the community. Obviously I over estimated how familiar people are with the fast and loose nature of TheInquirer. Next time I will preface anything from TheInquirer with a warning to always consider the source.

Sheesh, if it makes anyone feel better I'm still getting an XPS 1530 with an 8600GT.

 

Well, I just received my notebook back from HP today after I sent it in for repairs due to excessive heat (especially from the GPU - 109C).

Lo and behold:



This is from playing GW...Guild Wars...which is not very hardware intensive at all...

Fine, everything was maxed out. But I was playing in the first campaign when this occurred. The first campaign came out in 2005, and even an ATi Radeon X200 could run it. Medium Power Setting.

HP "repaired" my laptop, but I guess it's pretty hard to repair a defective product, isn't it?

If I keep this up my GPU will be dead in what, a year at most?

 

That 99c is just too damn hot. I'm just waiting for the national media outlets like CNN to pick up on this. Maybe at point, Nvidia, HP, Sony, Apple & Dell will let us know what's going on.

 

Yea, 99C is hot...

It hit 109C before it got "repaired", so I guess I'm in for some more weeks talking to HP guys...

At least now I got a Case Manager...

 

no it won't think about it.

lets say your gpu hit temps of 100+ degrees while gaming. Your fans will be already on max. turning the fans on max when it is lets say 70-80 degrees is not going to do anything. it'll slow down the heat but it's still gonna hit 100+ degrees if you are gaming more than a few minutes.

 

whoa.. 99c and 109c.. on my 17" mbp 2.6ghz model with the nvidia 8600m gt card with pushing it to the limit of what I usually do under osx.. i've only seen the gpu temp max out at 68C... and idles at 43-57 generally..

But I dont game nor do I use anything 3d intensive.. but with everything I normally do the gpu resides at 55-57c nothing higher. Only when I run virtual desktops and open up 30+ apps will it go to 68C but no higher.

My question is since overheating is not the cause of the gpu to die but the heat cycling that worries me even though my gpu never overheats anywhere more than 68C.

 

If you don't vsync it it will still run at 100% so that doesn't change anything. For me the fans just don't run on GW so I get like... 100C on power saver at 200/250 lol. Today AoC decided to stabilize at 106C with folding@home forcing my fans. That's crazy!

 

I just checked mine, I've got a 8600m gt from my 17" Dell Vostro mine is at a cool 55 c

that said, my 1yr warranty is up for renewal in two weeks, should I renew or hope that dell has a recall policy coming soon?

 

should i stay away from the 8400 gs? because i have a good deal on an m1330 with an 8400 in it and i do game (mostly just cs:s, with the ocasional rts, and gonna be getting left 4 dead for pc; the rest of my gaming is on my 360 but im getting off topic here ). i do have a notebook cooling pad which keeps my current laptop (specs in sig) temps down when gaming but im not sure by how much. i guess i can post the temps with/without the cooling pad maybe tonight if i decide to game a bit.

but my main question here is: should i still stay away from an m1330 with an 8400gs if i have a notebook cooling pad?

 

Holy p00p! That's some high temps. Max temp i got was 89C today.

If my fan kicks in at full at 85C but usually loads at around 83C then my fans aren't going to hit full speed until it gets hotter.

Think about it.

 

Well, other than some 'articles' on the net, and poor cooling system from some manufacturers (i.e. Alienware and others), is there any irrefutable proof that nvidia chips are faulty??

 

Apart from the fact that nvidia said they were. True, we have no ideas as to officially which ones may be faulty. Only when the appropriate notebook manufacturer tells us will we have irrefutable proof/evidence.

 

Take a look at nVidia's 8k form filing with the SEC.

http://biz.yahoo.com/e/080702/nvda8-k.html

You don't take a $200 million charge against earnings for repair and replacement costs if there is nothing wrong with your product.

 

Just want to make sure. The 9600 isnt considered a tweaked 8600 right?

 

Oh man, I just bought an m15x with the 8800M GTX on it. Ah well, let's hope they'll completely open up to their mistakes (or elaborate more on the faults on the chips) before the year of my warranty on this notebook expires.

 

The 8800M GTX is not affected, your laptop is fine.

 

Ah. I assumed that it did sinc-- wait a minute. Totally did not see that it was listed as G80. Haha sorry, silly me. Whew.

Anywho, even if the problems are only affecting G84 and G86's, how likely is it that we're just understating the faults that are at hand and that other cores are also faulty? Perhaps NV is trying to hide the rest of what hasn't been leaked to us yet.

 

I believe that I am one of the people that has a defective core. I have a Dell m1530 xps laptop, and after a couple weeks of gaming, performance got worse and worse. Comparing with other m1530 laptops, my GPU performs 10 to 20 degrees C higher at around 83 C (according to Rivatuner), the temperature where the GPU decides to start a cooling cycle and reduce gaming performance to around 10 fps.

i'm sure many people have gotten fully functioning GPU's, but I also believe a smaller minority like myself have been screwed. I don't think this news is overblown, because I may be one of the ones affected.

 

And exactly how do you know this is definatey the case. No one has "officially" stated which cores are effected. (Unless I'm missing something, that is.)

 

ever since i got my laptop back from HP (for GFX failure), the GFX temp hasnt gone over 60C (even while playing COD4)

 

What temperatures did it have prior to the failure?

 

If G92 were affected too, Nvidia should just declare bankrupt right now.

The 9650M GS is a rebranded 8700M GT so yes, it would affected as much as the 9500M GS
9600M GS, 9600M GT & 9700M GT should be fine since they are G96s

 

wow ati should come out with a bunch of new cards now. they could probably gein so much more market share considering all these nvidia problems as of late

 

I see a lot of people saying that since this is about last gen products that we should have seen problems LONG before now.....

I would like to add that it took 3 years for most bad capacitors to pop, following the electrolyte scandal, that was rumours and poohpoohed by many people in the two years before a massive bunch of motherboards and other components failed.

the other thing to consider my friends is, that nvidia have stated that they are investigating other products, so you could find G92 or even Nforce chipsets being affected.

the packaging material error, could be at a very base level, and little errors like that can happen to the biggest companies.

It only takes one employee to mess up big enough to take down a major player.

but of course we should wait for more information
but also consider this.... would nvidia really want to release more damaging information, or should they just replace the ones that fail within warranty,
as heat problems are most often a cause of gradual failure and performance degradation over time, which you would probably first blame on a tired windows installation.

just me 2c
lets hope this all blows over and everyone is actually fine.

 

is anyone experiencing a gradual degradation of performance over time? If so, do you have 3d mark scores to back this up?

 

Is the 8800M GTS affected?

 

I believe all nvidia laptop GPUs have a problem and its not just the G86/84. I say so because i have a Dell Inspiron E1705 that had a GeForce 7800 that died after lets say 20 months of use and the reason was heat issues. Now i had to extend the warranty on my E1705 to get Dell to replace the GPU. The new GPU is already in bad shape after 2 months of use. I see multi colored dots on the display when i play a video file and then games like FEAR which i had played earlier without any problems show all pixellated now. So i guess the card is pretty much dead i guess.

Now i bought a XPS M1330 for my wife and an XPS M1530 for myself recently. The 1330 shows temps at avg of 85C on GPU-Z, thats after i installed the latest BIOS and drivers from Dell. My 1530 seems to be doing better at an avg of 62C. It was at 78C before the driver update.
So thats three cards now that have pretty high temps. My E1705 with the 7800 always avg at 85C on idling.
Now i was readying forums of Alienware m15x and even they had heat issues earlier on, which were rectified after bios and driver updates. So i guess every nvidia GPU has a heating up issue. I hope nvidia come back with a solution cause i love nvidia and am neva going for an ATI.

 

Please read at least some of the thread. It has been stated many times- The 8800m GTS is built on the G92m core. As of now, it is not affected.

And to Loreto, the M15x was having problems due to it's build design. Most others with 8800m's don't run too hot, at least given the power. My 8800m GTS runs at 43C idle, and 67C full load, and that's with no cooling pad, just propped up from behind the battery.

 

are the new ones coming out fixed? I know it sounds like a stupid question but I'm laptop shopping for a friend and recommending a g84 but i want to make sure their fixed b4 i do

 

In short. No. You can't fix something when you don't know exactly whats wrong.

 

My 8600gt also reduces my gaming performance after a while, but its random. I never get over 81 C but sometimes the game would freeze for a while and the fps would go down to like 10-15 fps. I automatically downclocks. Anyway to fix this? Would sending this to Sager fix this or not?