DCS World 1.5 seems to have killed my graphics card

[SyntaxError]

Loaded up the qickstart SU27 intercept mission and managed to fly around for about 5 minutes when suddenly my PC and audio froze to a black screen. Upon reboot green dots and lines everywhere with no nvidia drivers loading. I put my GTX560 into another PC to test and same green artifacts. So not sure what happened but 1.5 killed my graphics card. My full PC specs:

 

AMD Phenom II X4 955BE

Gigabyte GTX 560 OC

DDR3 8GB 1333mhz

MSI 890GXM-G65

Windows 7 Ultimate

 

Just to note I have never over-clocked any of the components..

Edited October 2, 2015 by SyntaxError

[StandingCow]

What were your temps? Chances are that your card was near death or they tried to run at full and overheated. Poor case heating?

[Devrim]

...gigabyte gtx 560...

I don't know if it does matter Gigabyte or Asus, but my Asus nVidia GTX560 has worked perfect in higher settings (without DOF) all night.

[SyntaxError]

No need to turn this into asus vs gigabyte etc.. thread.

 

On-topic: Devrim & StandingCow I'm thinking maybe it was just an overheating issue caused by EDGE. Naturally I've been running this card for loads of games since I got the card and have never experienced any artifice or overheating issues. Either EDGE pushed my card OVER the limit and it overheated to the point of damaging the card, or the card was just getting old and EDGE was the last straw. Most times my card runs at the normal temperatures under-load for my climate at roughly 70°C +/-.

 

Sadly I was not monitoring my temps becuase I had stayed up an extra 5 hours to finish the download and was eager to run a test mission to see what performance I got.

 

I'm not sure if theres anything ED can do to investigate but I suggest they do if this can prevent this happening to other users (some users are already reporting artifacts like black textures and freezes to black screens).

 

Now I will have to convince myself to shelf out for a new GTX 970, PSU, and a new case in the meantime from my current life savings. :(

Edited October 2, 2015 by SyntaxError

[StandingCow]

No need to turn this into asus vs gigabyte etc.. thread.

 

On-topic: Devrim & StandingCow I'm thinking maybe it was just an overheating issue caused by EDGE. Naturally I've been running this card for loads of games since I got the card and have never experienced any artifice or overheating issues. Either EDGE pushed my card OVER the limit and it overheated to the point of damaging the card, or the card was just getting old and EDGE was the last straw. Most times my card runs at the normal temperatures under-load for my climate at roughly 70°C +/-.

 

Sadly I was not monitoring my temps becuase I had stayed up an extra 5 hours to finish the download and was eager to run a test mission to see what performance I got.

 

I'm not sure if theres anything ED can do to investigate but I suggest they do if this can prevent this happening to other users (some users are already reporting artifacts like black textures and freezes to black screens).

 

Now I will have to convince myself to shelf out for a new GTX 970, PSU, and a new case in the meantime from my current life savings. :(

 

Edge runs my cards (980's) near 100%, but that's fine because I have really good airflow in my case. What does your case/fan setup look like?

 

I have always said that games cannot kill cards due to overheating, instead they just expose poor cooling to the GPU and/or CPU.

[Boomer]

I always thought there was a "failsafe" that the system will shutdown after certain temperature is achieved, either on CPU or GPU. :huh:

 

edit:

 

yup, here it is:

 

http://nvidia.custhelp.com/app/answers/detail/a_id/2752/~/nvidia-gpu-maximum-operating-temperature-and-overheating

 

So, not quite sure what happened there.

Edited October 2, 2015 by Boomer

[SkateZilla]

Software doesnt kill GPUs,

 

GPUs fail due to previously dormant defect or heat damage from improper cooling.

 

Games due not manage GPU usage based on temps, that the driver and end users job.

[Brisse]

There's no way a game can kill your hardware. The drivers and hardware itself has failsafe features that will keep the temperatures in check at all times. It will even shut itself down if things are starting to get out of hand. This is the doing of either your graphics driver, or the hardware itself. You cannot blame DCS for this.

[SyntaxError]

I always thought there was a "failsafe" that the system will shutdown after certain temperature is achieved, either on CPU or GPU. :huh:

 

edit:

 

yup, here it is:

 

http://nvidia.custhelp.com/app/answers/detail/a_id/2752/~/nvidia-gpu-maximum-operating-temperature-and-overheating

 

So, not quite sure what happened there.

The PC did shut down but in the way of a bluescreen without the blue screen, just looped audio followed by total and complete darkness, so not anything like a CPU overheat shutdown. On a slightly related note I remember when StarCraft II launched there was a major fault with the main menu not having a frame-rate cap which led to quite a number of users with dead cards.

 

So I wonder did the overheat fail-safe fail or did it just get hot enough to expose/cause a fault in the card... I wish to damn I checked the temperature of the card when it was running but I was just too tired and just wanted to fly around for a bit.:crazy:

Edited October 2, 2015 by SyntaxError

[Brisse]

I know there was a driver problem that caused nvidia cards to die at some point earlier this year, or maybe last year but that was fixed long ago. You do have the latest drivers, do you not?

[SyntaxError]

I know there was a driver problem that caused nvidia cards to die at some point earlier this year, or maybe last year but that was fixed long ago. You do have the latest drivers, do you not?

Not at home right now to check but I believe it was 353.somethingsomething. Drivers and card always ran fine on most of the games that could put the GPU on full load.

[pr1malr8ge]

There's no way a game can kill your hardware. The drivers and hardware itself has failsafe features that will keep the temperatures in check at all times. It will even shut itself down if things are starting to get out of hand. This is the doing of either your graphics driver, or the hardware itself. You cannot blame DCS for this.

 

This is not true.. How ever it is unlikely.

A game if programmed maliciously or just happens to have unbeknownst "backdoor" that is affectingly overrides the driver to allow overheating can kill a card.

[SkateZilla]

He cards themselves have a temp shutoff, the only way to bypass it is to remove all cooling and allow heat to rise fast enough to fry the chip before the shutdown occurs.

 

heat can still destroy a card if left un managed over time.

[Viersbovsky]

It is quite possible that a game may destroy a GPU. For a prominent example, Starcracft2 was thought to do this (google it).

I had a game destroy my GPU some years ago, it was an alpha of some russian game that claimed Doom3 level graphics a full year before that game's release - it is not that uncommon.

Edited October 2, 2015 by Viersbovsky

[SyntaxError]

Just thought I'd post an update. I put my graphics card in my oven @ 200°C for 9 minutes and it fixed the problem.

[Nate--IRL--]

Hahah wow! That's an awesome fix :)

 

Nate

[gienkov]

Just thought I'd post an update. I put my graphics card in my oven @ 200°C for 9 minutes and it fixed the problem.

I did this 3 times by now and my 560 still works ;). Use lower temp next time, I use about 175 and 8 minutes only.

[tintifaxl]

The PC did shut down but in the way of a bluescreen without the blue screen, just looped audio followed by total and complete darkness, so not anything like a CPU overheat shutdown.

 

Call of Juarez did this to my then 18 mo. old AMD 7970. After reboot 2D worked fine, but starting a 3D app immediatly showed artifacts and after a few seconds a black screen.

 

HIS exchanged the card without questions asked. Maybe you could try to contact the manufacturer of the card - maybe there's still warranty to be had ...

[SyntaxError]

Just fired up 1.5 again and I'm not surprised my card couldn't handle it; it's running at 96°C at 98% load. I've setup the autoexec with a max fps of 60 just to be safe, but sadly since I can't move from my area to a colder one I think I just might have to pick up that GTX970. :noexpression:

 

At least baking my card was a fun experiment, but I can't quit DCS.

[Brisse]

Just fired up 1.5 again and I'm not surprised my card couldn't handle it; it's running at 96°C at 98% load

 

That doesn't mean the card is dead. Maybe you killed it in the oven, but try taking the cooler off, clean the area where heat flows from GPU to heatsink, apply new thermal paste and make sure you mount the cooler properly with a good amount of pressure against the GPU. Something may have warped when you put it in the oven, so that the GPU and heatsink doesn't have good contact anymore.

[SyntaxError]

That doesn't mean the card is dead. Maybe you killed it in the oven, but try taking the cooler off, clean the area where heat flows from GPU to heatsink, apply new thermal paste and make sure you mount the cooler properly with a good amount of pressure against the GPU. Something may have warped when you put it in the oven, so that the GPU and heatsink doesn't have good contact anymore.

No the card is working fine now after baking it. I'm saying the high temps are what damaged the card in the first place, and I don't think putting it through that much heat again will do anything good.
Edited October 6, 2015 by SyntaxError

[Brisse]

No the card is working fine now after baking it. I'm saying the high temps are what damaged the card in the first place, and I don't think putting it through that much heat again will do anything good.

 

So clearly you have a problem with the cooler, and yet you blame DCS for the damage done? Normal full load temperature for your card is 65-70°C. You are way beyond that, so it's easy to see where the problem is.

 

Baking it in the oven is not a proper fix. Reattach the cooler with clean surfaces, new thermal paste and make sure it's a tight fit, and not loose in any way. That's a proper fix.

[SkateZilla]

you likely baked the thermal paste and dried it out, and it needs replaced, check fan rpm and profile as well.

 

Couldve also bent the pcb or heatsink.

Edited October 6, 2015 by SkateZilla

[Saruman]

Well, if you make a game menu go to 1 fantastillion of FPS, it should be expected that temps get rising. Real thing is that you should use Vertical Sync in your graphic cards preferences, at a global level. There's really no need for the videocard to draw more frames than the screen frequency can accept: it only brings bad things like stutters and tearing.

 

That said, no way a game can kill a videocard, unless the user mistreated it because of poor ventilation in the case or savage overclocking.

[Brisse]

Well, if you make a game menu go to 1 fantastillion of FPS, it should be expected that temps get rising

 

True, it's unnecessary and will draw a lot of power which turns into heat, but the cooling should always be able to handle it anyway. If it can't then something is wrong with it.