Is my graphics card dying? (pic)

tylrdiablos

Distinguished
Apr 27, 2009
12
0
18,510
Around 2 weeks ago I noticed my frame in Left 4 Dead and World of Warcraft dropping at random intervals.

I tried old drivers and eventually Windows 7 (public RC) with the drivers listed on the nVidia site and then
with older drivers that had been modded with the mobility modder.

At one point I thought that alt-tabbing out, starting Task Manager and then returning to the game, had fixed
the problem (the framerate evened out, but in all honesty - this "solution" on worked once or twice)
but then I was hit by this:

3583318468_b8c067de8c_b.jpg


That corrupted display was constantly moving, kinda like an old static-y tv set.


My laptop is an Acer Aspire 6920G. (Specs here: http://www.insidehw.com/Reviews/Notebooks/Acer-Aspire-6920G.html )

Everest Home Edition says the CPU is running at 42 degrees C but there's nothing listed under GPU.
(It's just a blank dialog screen.)
There's nothing in the nVidia Control Panel for temperatures either.


Does anyone have any idea's on how to solve (or even troubleshoot) this problem?

Acer have said I can return the laptop but I'd still like to know if the problem is fixable.

I'm on the verge of buying a new desktop so a dead gfx card isn't a huge problem.


Thanks for your time, I hope someone has some idea's.
 
Used to have this problem with different games,after a restart it fixed itself, and i attributed it to my gfx running too hot (added cooling and it stopped).

ironically when this graphics display error happened with me it was when i minimized out of warcraft 3 :x
 
How do I go about adding cooling to a laptop (other than a desktop fan which I'll be picking up tomorrow.)

Are there any programs I can run to test the video RAM?


Thank you for your replies. Most helpful.
 
It does look like your video card is going bad. However, you can try downloading and installing OCCT. It should detect your GPU temperature. If it's really high, then you may simply need to clean out the dust from the cooling fan using a pressurized air can.

Furthermore, OCCT is able to test your video memory (amongst other things) as long as you have a DX10 compatible video card.
 
Ok, I've just ran OCCT's GPU: Memtest and it's reported 1280 errors.

The temperature box is disabled until I install the latest DirectX.

I'll do that now, restart machine and report back with a temperature.
 
Ok, the GPU Temperature test ran for 100 seconds in windowed mode.

The temperature started at 50 and was at 61 when the test was closed.

(If I remember rightly, anything over 75 is too much?)

Are we settling on overheating or bad GPU RAM?
 


Errors suggest that your video card's memory is faulty. It may have been damaged due to insufficient cooling over long periods of use (laptops can be prone to this if poorly designed/maintained). If your laptop is still under warranty, then you'll need to get it repaired or replaced.