AMD: R8xx Speculation

[Groo The Wanderer]

I don't see much of a general overheating problem there (ofcourse it can be for a few ppl with bad cooling), but several reports (especially on the first link) of ppl experiencing the crashes at low temperatures. Looks more like general problems (with the game/render code) in high settings than overheating. But ofcourse it's more conveniant for the authors to refer to the crashes as "the overheating problem"

The bad bumps failures are not 'crashes', when you reboot, they don't recover.

-Charlie

 

[Andrew Lauritzen]

The bad bumps failures are not 'crashes', when you reboot, they don't recover.

Furthermore WDDM can recover from the large majority of application/driver crashes nowadays... if it's hanging the machine something bad is happening outside the application.

 

[flynn]

Guys, I'm not a mod but can we get back to the topic, namely the R8xx series and do the NV talk in the other NV-related threads? The bumpgate thing has been discussed to death already anyway.

 

[Acert93]

Its funny you say that as there are several cases where people with enough now how have posted about "Adjusting", ie: bending the heatpipe a tad, the heatpipe/hs for the GPU in their laptops with Nvidia GPUs and noted a sizable decrease in GPU temps in both idle and heavy usage situations. But according to you, that isn't an OEM engineering problem. Seems to me, had the OEM insured proper HS/heatpipe installation on those GPUs, the numbers of problems would not have been so great.

And there are tens of thousands of cases of due to normal temperature changes chips up and dieing. There is a reason NV has been paying through the nose in regards to this issue. The OEM's sin was continueing to sell this CRAP and not disclosing it for months and months after they knew NV's product had an issue. Probably didn't help NV was blaming everyone else.

 

[Psycho]

The bad bumps failures are not 'crashes', when you reboot, they don't recover.

I was referring to Jawed's post about bloodball - ie. the "power virus" subthread, not bumpgate.

 

[Groo The Wanderer]

I was referring to Jawed's post about bloodball - ie. the "power virus" subthread, not bumpgate.

Ah, sorry.

-Charlie

 

[XMAN26]

If that is true, why do such incompetent engineering chuckleheads at HP. Dell, Apple, Lenovo, Sony, Acer, Toshiba and others only screw up on models with NV chips?

Also, these chips are running in spec, why do you need more cooling?

-Charlie

I wasn't saying Nvidia wasn't to blame brainchild, only that the OEMs also have to take some of the blame for this when a SIMPLE augmentation of the HPS(Heatsink/pipe) assembly can cause a conciderable decrease in idle and load temps on the GPUs that were affected. And you still never answered my question here or from when you were are theinq, why is this sololy Nvidias fault when the OEMs didn't insure proper cooling?

Yes they went by Nvidia specs for cooling, but when the OEM doesn't insure proper seating of the cooling unit to the GPU, how is that Nvidia's fault. It'd be like eVGA blaming Nvidia for failing GPUs because they(eVGA) designed a HSF unit to use instead of the Nvidia reference design that didn't sit flush on the GPU.

 

[neliz]

I wasn't saying Nvidia wasn't to blame brainchild, only that the OEMs also have to take some of the blame for this when a SIMPLE augmentation of the HPS(Heatsink/pipe) assembly can cause a conciderable decrease in idle and load temps on the GPUs that were affected. And you still never answered my question here or from when you were are theinq, why is this sololy Nvidias fault when the OEMs didn't insure proper cooling?

Yes they went by Nvidia specs for cooling, but when the OEM doesn't insure proper seating of the cooling unit to the GPU, how is that Nvidia's fault. It'd be like eVGA blaming Nvidia for failing GPUs because they(eVGA) designed a HSF unit to use instead of the Nvidia reference design that didn't sit flush on the GPU.

they either get or ask for a GPU within a thermal spec. nV gave them one that can't cope.

 

[Groo The Wanderer]

I wasn't saying Nvidia wasn't to blame brainchild, only that the OEMs also have to take some of the blame for this when a SIMPLE augmentation of the HPS(Heatsink/pipe) assembly can cause a conciderable decrease in idle and load temps on the GPUs that were affected. And you still never answered my question here or from when you were are theinq, why is this sololy Nvidias fault when the OEMs didn't insure proper cooling?

Yes they went by Nvidia specs for cooling, but when the OEM doesn't insure proper seating of the cooling unit to the GPU, how is that Nvidia's fault. It'd be like eVGA blaming Nvidia for failing GPUs because they(eVGA) designed a HSF unit to use instead of the Nvidia reference design that didn't sit flush on the GPU.

Wow, you are getting desperate. Let me ask you another question, why is it that only Nvidia laptops, across 10 or so manufacturers, from NV42 to G9x, have ill fitting heatsinks?

Do you think it is because the overflowing awsumn3ss of the NV chips push the heatsink out, or do you think that is the way they all come, and that is 'spec'? Anyone can do better than spec with a little TLC, that is the bane of mass production.

I look forward to you strained explanation of why only NV based parts have badly attached thermal solutions.

-Charlie

 

[Silent_Buddha]

I wasn't saying Nvidia wasn't to blame brainchild, only that the OEMs also have to take some of the blame for this when a SIMPLE augmentation of the HPS(Heatsink/pipe) assembly can cause a conciderable decrease in idle and load temps on the GPUs that were affected. And you still never answered my question here or from when you were are theinq, why is this sololy Nvidias fault when the OEMs didn't insure proper cooling?

Yes they went by Nvidia specs for cooling, but when the OEM doesn't insure proper seating of the cooling unit to the GPU, how is that Nvidia's fault. It'd be like eVGA blaming Nvidia for failing GPUs because they(eVGA) designed a HSF unit to use instead of the Nvidia reference design that didn't sit flush on the GPU.

The problem here is Nvidia TOLD them the specs for the chip. Nvidia TOLD them the "safe" operating temperatures. Nvidia TOLD them what the chips should run at.

Standard operating procedure for any chips used by any OEM.

OEM's then design to specificiations provided by the IHVs. In this case they designed the cooling to the specs provided by Nvidia.

I'm not sure how people can get off blaming OEMs on this one. If it was just one OEM, sure they probably screwed up. Multiple OEMs purposely slitting their financial wrists designing out of spec cooling? Uh, yeah...

And BTW - OEM's also tend to test their cooling setups AFTER integration to make sure the chips are running at spec (in this case provided by Nvidia) in fully configured and shipping laptops. Again, standard operating procedure.

And as someone mentioned previously, if there was systemic incompetence at ALL major OEMs, this would have affected notebooks with other chips from both Nvidia and AMD. Only that wasn't the case.

Regards,
SB

 

[XMAN26]

I'm not saying Nvidia isn't to blame, just not all the blame lays on them. And FYI, since being at work the last 2 days, I did some checking on the laptops with Nvidia GPUs in them. 3 Dell, 2 HP, 4 Comcrap and 1 Lenovo. Out of all of them in for service, only 1, ONE, UNO should signs on the heatsink/pipe assembly with a clean true(100%) seating to the GPU and that was the Lenovo. They use this really cool thermal pad that is cold to the touch. All the others had missing contact of atmost 50% of the GPU. Somehow I doubt we are that unlucky as a shop that all the other makes showed no 100% contact at all according to all you there should be. We also have 4 laptop with ATI GPUs which failed. All four less than 18m old and also showed less than 100% contact.

So again, I ask, why is it ALL Nvidia's fault for this mess? Again, it would be akin to blaming Nvidia for a cooler an AIB partner designed and used that caused GPU failure because it didn'tsit flat on the GPU. The AIB has the specs for the GPU and knows the thermal limits.

Anyone could design, make, test and sell a CPU or GPU that has a thermal design in mind, but if the OEMs you are selling to dont ensure proper HS/P seating, how in the hell would that be any one snigle companies fault? The blame would fall and should fall on atleast 2 companies. The chip maker for not playing it safe with therms in laptops and teh OEMs for shoddy builds and poor QA.

The problem here is Nvidia TOLD them the specs for the chip. Nvidia TOLD them the "safe" operating temperatures. Nvidia TOLD them what the chips should run at.

Standard operating procedure for any chips used by any OEM.

OEM's then design to specificiations provided by the IHVs. In this case they designed the cooling to the specs provided by Nvidia.

I'm not sure how people can get off blaming OEMs on this one. If it was just one OEM, sure they probably screwed up. Multiple OEMs purposely slitting their financial wrists designing out of spec cooling? Uh, yeah...

And BTW - OEM's also tend to test their cooling setups AFTER integration to make sure the chips are running at spec (in this case provided by Nvidia) in fully configured and shipping laptops. Again, standard operating procedure.

And as someone mentioned previously, if there was systemic incompetence at ALL major OEMs, this would have affected notebooks with other chips from both Nvidia and AMD. Only that wasn't the case.

Regards,
SB

And the therms Nvidia said were safe I'm sure were based on 100% HS/P assembly contact with the GPU, not partial. Please, SB, Groo "I make up half of what I report about Nvidia" Dem and no-x how any GPU is expected to last for any period of time at high temps without proper HS/P assembly seating.

And SB, it myay be SoP to test the setup, but do you honestly think any of the OEMs treat the laptops like most people do? Play heavy gaming for 3-4 hours, turn off, start up repeat, turn off. Do some homework, surfing, light work, sleep/standby/hibernate, back to heavy gaming. Turn it off. ALso while also do all this sometimes using it in their laps, on pillows, beds, blankets, over top loose papers for month on end? I some how seriously doubt they stress test their laptops anywhere near that strenuous.

 

[LordEC911]

So then why is Nvidia paying out money if it isn't their fault?
Why isn't it covered by their insurance?

You seem to not want to believe the facts that everyone is laying out in front of you...

 

[dizietsma]

I was speculating yesterday whether buying a dual gpu R800 would be "the last video card I would ever need to buy" so to speak. I could imagine it might have a very long lifetime compared to previous generations due to all this talk of PC games no longer advancing as fast due to concentration on consoles. Also I play at 1920x1200 on a 24 inch monitor so it's not as if I am pushing the pixel numbers either.

Maybe, for once, the top dog expensive card is pretty much futureproof, at least as far as my setup.

 

[XMAN26]

So then why is Nvidia paying out money if it isn't their fault?
Why isn't it covered by their insurance?

You seem to not want to believe the facts that everyone is laying out in front of you...

Readnig comprehension 101, "I'm not saying Nvidia isn't to blame," My opening statement. They are to blame and should pay. They knew what the issue was, tried to bury it, then denied it. But the OEMs share in the blame for the falures. 8600GTs were supposedly affected by this in desktops but yet they haven't failed by the droves. Could this be because the HSF assembly ACTUALLY MAKES 100% CONTACT WITH THE GPU? I dont know about you, but when the same part in two different applications has only extremely bad failure rates in one fo the two applications it is being used in, it seems to me it is more than misguided therm and cooling specs from Nvidia and the substrate.

 

[compres]

Readnig comprehension 101, "I'm not saying Nvidia isn't to blame," My opening statement. They are to blame and should pay. They knew what the issue was, tried to bury it, then denied it. But the OEMs share in the blame for the falures. 8600GTs were supposedly affected by this in desktops but yet they haven't failed by the droves. Could this be because the HSF assembly ACTUALLY MAKES 100% CONTACT WITH THE GPU? I dont know about you, but when the same part in two different applications has only extremely bad failure rates in one fo the two applications it is being used in, it seems to me it is more than misguided therm and cooling specs from Nvidia and the substrate.

Still does not explain why several manufacturers had the problem.
Still does not explain why ATI chips did not have the same problem.

 

[XMAN26]

Still does not explain why several manufacturers had the problem.
Still does not explain why ATI chips did not have the same problem.

Missed this didn't you from an earlier post by me:

"I did some checking on the laptops with Nvidia GPUs in them. 3 Dell, 2 HP, 4 Comcrap and 1 Lenovo. Out of all of them in for service, only 1, ONE, UNO should signs on the heatsink/pipe assembly with a clean true(100%) seating to the GPU and that was the Lenovo. They use this really cool thermal pad that is cold to the touch. All the others had missing contact of atmost 50% of the GPU."

Add that to the manufactoring snafu and you get massive failures of GPUs in laptops. Still funny how desktop 8600GT/GTS don't seem to suffer from this same issue in any where near the same numbers. Proper cooling does alot.

 

[Psycho]

Notebooks have much longer design cycles than which HSF to slam on a new graphics card. So the notebook cooling was probably designed for the specified TDP and not the higher actual TDP. IIRC this is part of the story.
On top notebook gfx generally get runs more hot / closer to the specified limits, as it is more difficult (/noisy etc) to cool them. So no wonder they fail first.

 

[Groo The Wanderer]

I'm not saying Nvidia isn't to blame, just not all the blame lays on them. And FYI, since being at work the last 2 days, I did some checking on the laptops with Nvidia GPUs in them. 3 Dell, 2 HP, 4 Comcrap and 1 Lenovo. Out of all of them in for service, only 1, ONE, UNO should signs on the heatsink/pipe assembly with a clean true(100%) seating to the GPU and that was the Lenovo. They use this really cool thermal pad that is cold to the touch. All the others had missing contact of atmost 50% of the GPU. Somehow I doubt we are that unlucky as a shop that all the other makes showed no 100% contact at all according to all you there should be. We also have 4 laptop with ATI GPUs which failed. All four less than 18m old and also showed less than 100% contact.

http://www.semiaccurate.com/2009/07/23/how-peek-chip-guts-without-killing-them/

Please explain the third picture, that is a G200 with lots of missing TIM under the lid. Does this mean NV is incompetent? That would be a yes, but I do look forward to hearing your opinion.

So again, I ask, why is it ALL Nvidia's fault for this mess? Again, it would be akin to blaming Nvidia for a cooler an AIB partner designed and used that caused GPU failure because it didn'tsit flat on the GPU. The AIB has the specs for the GPU and knows the thermal limits.

OK, so if all these AIBs are incompetent, why are they only incompetent making NV laptops? Why don't ATI parts suffer from the same catastrophic failure rates as NV ones do, your rather biased, if real, sub-sub-sub-sample not withstanding.

The part that you don't understand is that the half-connected parts are indeed running in spec. As long as they maintain a temperature as specified by NV, they are in spec. The chips maintain that temp.

Look at the graph here:
http://www.theinquirer.net/inquirer/news/1013947/nvidia-should-defective-chips
Look at the recommended temps for the GPUs that you are talking about. See a problem? The GPUs are failing within the recommended temperature range, overheating has nothing to do with it. THIS IS NOT AN OVERHEATING PROBLEM, it is an incompetent engineering problem at the chip level, not the laptop level.

Any heatsink attach problem is not an issue here.

Anyone could design, make, test and sell a CPU or GPU that has a thermal design in mind, but if the OEMs you are selling to dont ensure proper HS/P seating, how in the hell would that be any one snigle companies fault? The blame would fall and should fall on atleast 2 companies. The chip maker for not playing it safe with therms in laptops and teh OEMs for shoddy builds and poor QA.

It doesn't matter. If you do believe it matters, then you really need to answer the question of why 10+ OEMs failed to engineer proper cooling solutions only on their NV based models. In fact, on some, where there was a choice via MCM modules, only the NV parts had 'faulty thermals', and only they failed. Hmmm.....

And the therms Nvidia said were safe I'm sure were based on 100% HS/P assembly contact with the GPU, not partial. Please, SB, Groo "I make up half of what I report about Nvidia" Dem and no-x how any GPU is expected to last for any period of time at high temps without proper HS/P assembly seating.

No, they were not, and if you think they are, you are an idiot. There is _NO_ thermal solution specified in the design guides, only temps to remain in. If it is what you say it is, why did Nvidia design such a crappy thermal solution.

You need to keep a GPU in a certain temp range, and the OEMs did. It doesn't matter how they do it, with an aluminum slug, a vapor chamber, or fairy dust, the temp matters, and only the temp matters.

And SB, it myay be SoP to test the setup, but do you honestly think any of the OEMs treat the laptops like most people do? Play heavy gaming for 3-4 hours, turn off, start up repeat, turn off. Do some homework, surfing, light work, sleep/standby/hibernate, back to heavy gaming. Turn it off. ALso while also do all this sometimes using it in their laps, on pillows, beds, blankets, over top loose papers for month on end? I some how seriously doubt they stress test their laptops anywhere near that strenuous.

Yeah, actually I do think they treat it that way, and far worse. Having just spent a week in Taipei discussing thermal solutions for laptops at a conference, I do actually know how they test, and have been to several of their labs. You are dead wrong here.

So, to conclude, you seem to not want to answer the question about why only NV laptops have badly designed and connected HSFs. Why is it?

-Charlie

 

[Groo The Wanderer]

Missed this didn't you from an earlier post by me:

"I did some checking on the laptops with Nvidia GPUs in them. 3 Dell, 2 HP, 4 Comcrap and 1 Lenovo. Out of all of them in for service, only 1, ONE, UNO should signs on the heatsink/pipe assembly with a clean true(100%) seating to the GPU and that was the Lenovo. They use this really cool thermal pad that is cold to the touch. All the others had missing contact of atmost 50% of the GPU."

Add that to the manufactoring snafu and you get massive failures of GPUs in laptops. Still funny how desktop 8600GT/GTS don't seem to suffer from this same issue in any where near the same numbers. Proper cooling does alot.

He doesn't appear to have missed it, you just seem to be parroting back irrelevant info to dodge the question.

-Charlie

 

[Groo The Wanderer]

Notebooks have much longer design cycles than which HSF to slam on a new graphics card. So the notebook cooling was probably designed for the specified TDP and not the higher actual TDP. IIRC this is part of the story.
On top notebook gfx generally get runs more hot / closer to the specified limits, as it is more difficult (/noisy etc) to cool them. So no wonder they fail first.

I really doubt this. If the HSF was underspecced, you would get thermal runaway/overheating. In this case, you are not getting that, the failures occur at normal temps.

-Charlie