Sounds temp related. Maybe download MSI Afterburner and set it up with a aggressive fan profile for gaming sessions and a regular profile for net surfing. If you have EVGAs Precision installed, uninstall it and Nvidias Geforce Experience also if that's installed too. Yes, you can mention any brand you like here. :)

Since your issue doesn't appear to be temp related, and if it's not a power issue, maybe it's driver related...
 
I would run DDU in safe mode to clean out all your driver files > https://www.guru3d.com/files-details/display-driver-uninstaller-download.html
Then install the latest driver from Nvidia and see if that fixes your issue > http://www.nvidia.com/Download/index.aspx?lang=en-us
 

Thanks guys.
 
I did all kinds of fresh drivers installation including using DDU in Safe Mode. The problem existed in official drivers 388.13 until 388.71. I stopped updating drivers after since they did not provide extra game support I want and since GTX 680 is like seven years old now I thought further drivers wont make a difference unless they provide support for new games I play (or if I get those games in the first place).
 
I don't have another fully functional and software ready PC (games are annoying to transfer to another PC) so I just borrowed its PSU and installed it on the main PC. I tested it for some time and nothing happened yet but not extensive enough to tell. I'll test for couple more days and see.
 
Is there a chance it's PSU related?

The games I play that gave this problem are Dark Souls 3, Nioh and Gears of War 4. They are the only games I played since getting this card, really. It happens also in clam situations and almost no other players around. Nioh seemed to have the least problem occurrence and only after extended playing time.
 
I have a fair share of experience in detecting hardware defects but in whole parts only. Motherboard is the most difficult for me to detect and Sajin mentioned this possibility. I did use the card in two different PCI-E slots (the third cannot be used due to spacing) and the same problem happened.
 
All (at least basic ones) settings in BIOS were tested in fixed and automatic settings and within specs. The RAM voltage for example is set to the number on the sticker (I think it was 1.65v) and the it's an true 1600MHz (as in without XMP). It's new to me that such setting could crash this specific video cards this way but not the other one. I did try different BIOS settings including disabling all features I can find even turbo-boost and speed-step.
 
I made this thread having in mind thinking out of the box too along side obvious causes. Could it be, like, a damaged VRAM segment or something in the chip that can be taken care of somehow? Is there a professional way to detects such defects?
 
Dunno, but I have a strange feeling the least expected cause mentioned so far is the culprit; the PSU. It's wishful thinking, but you know, despair
 
And BTW, cool post count, wmmills!

Some of the older cards might have a longer warranty. If you properly registered the card when first purchased check the expiration date of the warranty by going to MEMBER at the top of this webpage and choose MY PRODUCTS

Smart_Guy
I believe this decides no warranty for me.
 
Can a high +12v current cause this?
 
Could it be the VRAM?
 
Come to think of it, can it be fixed if sent to EVGA?

Your card is definitely out of warranty at this point in time. 
 
First question: Overcurrent to the card could cause all sorts of issues. The card should shutdown though as it has ocp protection. Having a high amp rating on the 12v isn't the issue.
Second question: Yes, especially since you've been running with it underclocked.
Third question: No, evga would only replace the card with another card.
 

If your video card is indeed defect EVGA will probably not be able to repair it should it be out of warranty. You can ask EVGA tech support if you wish to have any additional info.

Thank you for the reply, Highiron.
 
Did you have a similar problem of video freezing but audio still playing?
 
I'll give that driver a try, at any rate. If that's the real cause, It would be unfortunate because newer games won't be fully (or at all) supported. But hey, beggar can't be choosers, right? (at least it's gonna work.)
 
I have an update tho. It's 100% certainly and definitely not a PSU problem. The crash/freeze happened even with the borrowed PSU. At least this is one thing I'm sure of. The main PSU is 1000W and it would be bad if it was faulty.

highiron

Smart_Guy
Did you have a similar problem of video freezing but audio still playing?

I had worse. I was getting a total computer freeze with the sound going wonky, not looping but playing in slow motion is the best I can describe. After about 20 seconds it would BSOD.
You can check the GTX 690 WATCHDOG violation thread below yours for more details. I wouldn't worry about about games being supported. In most cases it's just minor tweeks and Shadow Play and things like that. I have never encountered a game where an older driver doesn't work unless the driver is really old. Some newer drivers do more harm than good and can actually cause things to perform worse, such is my case at the moment. It would be accurate to say that any drivers released today are not updated with the 600 series cards in mind so getting the latest drivers may actually cause more problems than it actually fixes if there was anything to fix at all. The 385.69 driver is only about 6 months old so it's not like your going back 5 years or something like that.
 
Cheers.

I'm using a multi monitor setup so frozen games' tasks can be easily ended having the Task Manager on a secondary screen. I suppose some crashes probably don't allow anything else to come on top even the Windows logging menu so it could be seen as a full system crash. But that's not important now.
 
I got this card to SLI it with an existing one. I hope 385.69 at least has an efficient Dark Souls 3 SLI profile. Or can those be ripped from newer drivers and installed on older ones? I'll look that up. I'm waiting for Nioh's and AC: Origin's SLi profiles so yeah.
 
If it is a driver issue, then this is outrageous. Even the GTX 480 on my other setup works great on latest drivers (at least it still supports DX12).
 
I didn't know about that dump file. I'll see if I'm getting the same violation.
 
Edit:
Ha, looks like this card is considered too old. Registering the product says serial number not found.

Thanks for the tip. I get it is just a profile with some script that has nothing to do with (or require) something in the core design of drivers then. That's cool.

Smart_Guy
No use. Same problem happens with the drivers 385.69.

Sorry to hear that. It was a simple test so at least you can rule out the drivers.
 
If the card fails in a different build or a clean Windows install, indeed the card is likely faulty. Best of luck.
 
Cheers.

I don't know...it's coolermaster....done some lower end units...but about everyone has...got a couple low end units a corsair and a Apevia and everything seems okay..or something wrong with the card.

Update:
 
I tested the card on another computer for +10 hours and no crash (vs 6 max on the main computer). This means something is up with my main computer. If so and the card is fine, the question would be why would the other GTX 680 (the MSI) not get the same problem? Shouldn't both be basically the same? Would the MSI's BIOS flashed on the EVGA card make it treated as one?
 
There are some details I think I should have provided too; I'm on a multi monitor setup and ever since adding the second card maximized windows on the main screen refuse to receive any input after quitting a full screen game until minimized from the task bar. The multi monitor setup is a main 60" Smart TV and a secondary 27" Samsung monitor. The other test computer was a single monitor setup.
 
Next test is with turning off multi monitor function since that also requires grafix processing maybe. Others will follow.

Smart_Guy
Update:
 
I tested the card on another computer for +10 hours and no crash (vs 6 max on the main computer). This means something is up with my main computer.

Sounds like it's time to build a new rig.

Update:
Unbelievable! See how much time has passed since I said the problem stopped? Now it suddenly came back and probably worse! I didn't test it on GoW4 tho, I just kept playing DS3, and it was not just 1-2 hours a day. Comparing the two test computers, the main uses a video smoothing software called SVP that utilizes OpenCL. Out of despair, I disabled it and crashes stopped. Could be luck and the card decided to not crash so far. Let's see how it goes.