Black Octagon
Golden Member
- Dec 10, 2012
- 1,410
- 2
- 81
toyota
Lifer
- Apr 15, 2001
- 12,957
- 1
- 0
how many times does it have to be said that the driver was not letting the protection kick in?
Depends entirely on the hardware.
Most reasonably powerful hardware can be damaged or killed with high enough temperatures generated through its own operation. A lot of modern hardware has built in hardware protection that'll literally kill operation of the device at certain trigger thresholds.
There have been driver releases that effect fan control and cause dangerous temps but theoretically that can only kill or damage cards which lack hardware based thermal protection.
Most reasonably powerful hardware can be damaged or killed with high enough temperatures generated through its own operation. A lot of modern hardware has built in hardware protection that'll literally kill operation of the device at certain trigger thresholds.
There have been driver releases that effect fan control and cause dangerous temps but theoretically that can only kill or damage cards which lack hardware based thermal protection.
- Sep 28, 2005
- 21,077
- 3,578
- 126
I have yet to see 1 company that has protection on VRMS.
They always assume... if the cpu throttles... vrms wont matter... however..
VRMS are also heat sensitive, and on modern gpu's they require cooling.
VRMS get Toasty... which is why a lot of aftermarket sinks FAILED without VRM sinks.
Once these VRMS go on a runaway heat ramp... it leads to PCB dmg... degradation of card and so on.
i think 3... however toyota what did u mean again?
(sarcasm)
Boiling water = 100C!
i think not... cuz ur tubing would of gone YAMYAMYAM from that heat.
You would of formed vapor bubbles which would of increased the internal pressure on your entire loop.
Your tubing would of exploded off the gpu block regardless of clamps from the internal pressure and the tubing being YAMYAMYAM.
(YAMYAMYAM = VERY VERY SOFT... boil a piece of tubing and u'll understand....)
i think more likely ur water was undergoing convection, cold water being replaced with warm water which was rising.
You probably never exceeded 60C coolant side... however your gpu could of been over 100C.
And if ur running a FC block, that's still enough protection for the VRMs versus a unprotected air sink fan which has profiles wrong causing the sink to undergo runaway heat.
Runaway heat -> when your cooling potential cant keep up with heat load... so the heat compounds and compounds until equilibrium is met.
All sinks in general are more efficient at higher temps... most of the time the sweet spot in efficiency tho isnt where we want it to be.
In coppers case, it becomes more efficient as u get further from ambient.
Ive melted GPU blocks BTW balla, that scenario im telling u was mine.... in cases like that... the card was DEAD.
Mine looked exactly like this guys did.
Boiling water inside GPU block => NO FUN AT ALL!!!
Last edited:
@ aigomorla at the time I had the cap off my reservoir, I can remember it plan as day...
I was playing Diablo 3 hardcore of course, and all of a sudden out of the blue my system just shut off. So I turned it back on, fired up the game and looked over at my OSD and my gpu was reading 103C, then it shut back down...
I looked over at my setup and the hoses were drooping down, they were so soft they weren't retaining their form any longer. Outside the the GPU block the rubber had started to stertch and there was a gap of air which wasn't there before... There were bubbles coming from the block, and the area of hose without water was steamed.
I have no temp reading for the water; so I can't say what the water temp was though it was hot. I would say it was close to 160F, based on the tubing, the condensation, and the bubbles.
I say this because before water reaches the boiling point you see the formation of bubbles, which is what I believe I was seeing. So not 212F, lower, much lower, but still hot, very, very hot.
I was playing Diablo 3 hardcore of course, and all of a sudden out of the blue my system just shut off. So I turned it back on, fired up the game and looked over at my OSD and my gpu was reading 103C, then it shut back down...
I looked over at my setup and the hoses were drooping down, they were so soft they weren't retaining their form any longer. Outside the the GPU block the rubber had started to stertch and there was a gap of air which wasn't there before... There were bubbles coming from the block, and the area of hose without water was steamed.
I have no temp reading for the water; so I can't say what the water temp was though it was hot. I would say it was close to 160F, based on the tubing, the condensation, and the bubbles.
I say this because before water reaches the boiling point you see the formation of bubbles, which is what I believe I was seeing. So not 212F, lower, much lower, but still hot, very, very hot.
Last edited:
Gikaseixas
Platinum Member
- Jul 1, 2004
- 2,836
- 218
- 106
OMG don't remind me. The 470s were the hottest video cards I've ever used in my entire life; especially in SLI. After overclocking, the top card would routinely hit 100c, or even higher (I think the highest I saw was 102). The bottom card would usually be at 95 or 96c if I remember..
At the first opportunity, I used the EVGA step up program to get a pair of 480s and they ran much cooler due to those massive heat pipes.
Anyway, after reading the comments, I'm now convinced. I'm still shocked that drivers can affect so critical a hardware function though. I mean, drivers can easily become corrupted for a host of reasons, so it seems foolish to let them affect fan operation to such a degree..
You're just not thinking about this right. Think about it more like how the BIOS works on your PC.
Say you're installing Windows for the first time and you're using an adapter that needs a driver like a RAID card. BIOS can boot from it fine, but as soon as the Windows kernel dumps the BIOS interrupts bam no more RAID.
Same thing here. The card might have a fan program loaded in its BIOS. Driver loads, and the first thing it does is dump the fan program for its own.
Black Octagon
Golden Member
- Dec 10, 2012
- 1,410
- 2
- 81
That's the 2010 driver. The OP said that this thread is not about that one but rather the more recent 320.18 allegations. Your post seemed to be suggesting that NVIDIA had admitted that 320.18 was frying cards, and it is that particular idea that is lacking a hard source.
No, the OP made this thread because he couldn't believe Nvidia would ever make a bad driver. It starts at the end of the locked frame pacing thread.
I had 2x SLI GTX260s fail using 196.75
I was playing a game for around 2 hours or so when my PC suddenly cut out. Both cards overheated because the fans were not ramping up high enough to counter the heat.
http://www.hardocp.com/news/2010/03/05/nvidia_19675_drivers_kill_your_gpu
I was playing a game for around 2 hours or so when my PC suddenly cut out. Both cards overheated because the fans were not ramping up high enough to counter the heat.
http://www.hardocp.com/news/2010/03/05/nvidia_19675_drivers_kill_your_gpu
BrightCandle
Diamond Member
- Mar 15, 2007
- 4,762
- 0
- 76
When the 196.75 driver complaints started coming they really poured in, there were a lot of people all claiming it within a very short time after the driver was released. We have seen the same claim on every new Nvidia driver since, but the volumes of complaints make it unlikely that this is the true cause. Anything since 196.75 however hasn't been proved to be the driver and its unlikely to be so based on the volumes of complaints we have seen. But we can see the harm NVidia did to their reputation with that driver when 50 driver releases later people are still pointing at the latest ones and saying it broke my card in this way.
No doubt that was a very bad moment for NVidia but IMO not its lowest point, I view the systematic cheating on benchmarks to be worse as it was a calculated strategy of deception not a simple error that got passed testing.
No doubt that was a very bad moment for NVidia but IMO not its lowest point, I view the systematic cheating on benchmarks to be worse as it was a calculated strategy of deception not a simple error that got passed testing.
An00bis
Member
- Oct 6, 2012
- 82
- 0
- 0
I'm tired of seeing kids who are being faboys to certain hardware companies, monopoly is BAD, just look at cpus, if AMD was a little stronger intel wouldn't release K cpus at 50% higher price (it is like that in my country), and while you praise and praise Intel, you make Intel so cocky of itself they bribe many retailers to only include intel hardware and nvidia (not because they love nvidia, but because it at least hurts amd indirectly). I know a guy from another country, he asked me if i can help him make a PC and when I browsed some of the online retailers in his country i didn't see a trace of AMD, intel everywhere, and rarely nvidia, and some very old nvidias like the 8xxx series, 2xx, 4xx and few gtx650s. Too see how bad monopoly is, it's bad when any company does it, just like how overpriced amd set its cards at the beginning of the 7xxx series, now by the end of it most of its cards were reduced in price by half some more some less. I think I got my 7850 with about 300$, now it's 150$, monopoly is bad it doesn't matter whoever does it and you're simply an idiot if you become a fanboy to a hardware company just beacuse you either hate green or you hate red, or you liked a fancy gimmick (tressfx or physx) or fell for a stupid Ad/ While you fanboy and give them free publicity they jew out more money, beat the competition harder, are more confident in setting higher prices and stop wanting to improve their crappy hardware because they see no reason if the competition does nothing to challenge them. Remember how nvidia released some miraculous drivers right after amd published its "30% more performance in BF3" drivers? didn't it make you wonder? either nvidia was hoarding these fixes and didn't want to publish them, or its driver devs were visited by jesus and he told them every bit of line of code they had to implement or fix to make their gpus perform faster, and all this within a week or so since they saw amd publish those drivers.
Black Octagon
Golden Member
- Dec 10, 2012
- 1,410
- 2
- 81
http://forums.anandtech.com/showthread.php?t=2055899
It was real. It was discussed in this forum.
There shouldn't really be any debate about it.
It was real. It was discussed in this forum.
There shouldn't really be any debate about it.
Last edited:
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 23K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter