GPU Seems to be causing bsod

Killedan9

Honorable
Feb 26, 2013
10
0
10,510
So starting yesterday my computer has started blue screening when ever I play a game or run a graphics benchmark. Just prior to crashing the screen will start flickering and freeze up looking like this.

Through my testing I'm pretty sure that the graphics card is responsible but before I go out and buy a new one (pretty old, warranty is probably expired), I was hoping to get another opinion.

Resources:
Most recent memory dump
Second most recent memory dump
msinfo file

Specs:
Intel Core i7-4770K 3.5GHz Quad-Core Processor
G.Skill Ripjaws X Series 32GB (4 x 8GB) DDR3-1866 Memory
Asus GeForce GTX 780 3GB Video Card
Thermaltake EVO-Blue 850W PSU
MSI Z87-G45 Gaming ATX LGA1150 Motherboard

Symptoms:
The error messages that I received on the bsod screen varied quite a bit but the most frequent one was "Video_tdr_failure (nvlddmkm.sys)".
I also got IRQL_NOT_LESS_OR_EQUAL, VIDEO_SCHEDULER_INTERNAL_ERROR and sometimes just black screens with no shutdown.

Things I've tried (in rough order):
Uninstalled all graphics drivers.
Reset main drive to system image made about 8 months ago
Run memtestx86 for 3 passes
Reseated graphics card
Removed graphics card and run off of integrated graphics - WORKED!
 
Solution
You might have a degraded GPU. Try bringing your GPU voltage up a little with out overclocking. I would not go much over 1.2 volts. Let us know the results.

Killedan9

Honorable
Feb 26, 2013
10
0
10,510


I've run IntelBurn Test on max for half an hour with no problems to test the cpu
and as stated in the post I ran memtest86 overnight also with no problems
however when I run unigine valley bench mark with my graphics card plugged in the system will crash almost instantly.
when I run unigine without my graphics card running on integrated graphics it runs fine (Albeit alot slower)
 
You are indeed right, it is the graphics card.

Ok down to the causes of why you are seeing the awful screen discolouration and pink squares.

The first is obvious in most cases it is an indication that the Gpu on the card is getting too hot, or the Gddr 3 memory chips on the card also.

So check the temps. take the card out and check for any dust build up that may be blocking air flow from the fans through the fitted cooling solution on the graphics card.


2. Other causes are a lack of power being provided to the graphics card, meaning you should at least if you can try the system with another power supply of the same, or more wattage.

3. Failing capacitors on the graphics card its self, memory and display out put can be effected by this until the faulty capacitor is changed.

4.Memory chips that have a bad or cracked solder point on the underside of the chip where it contacts with the circuit board of the card. A re flow may be required.

5. The card is over clocked on the Gpu and the memory speeds.
Reduce the Gpu over clock, and the memory over clock on the card by 100Mhz each.

Then test the system to see if the screen corruption is resolved.


 

Killedan9

Honorable
Feb 26, 2013
10
0
10,510


So I just gave my gpu a good clean and put it back in my PC. When starting the pc up it took a few tries to get to windows because I was seeing the same distorted pixels and discoloration while on bios too.

I checked my overclock settings and everything was running standard (no overclock) and my temp was running at 35C while on the desktop.
The moment I started up the unigine valley benchmark the screen will distort and and then freeze (No bsod just stuck there). As far as i can tell the temp was never above 40C.

After this I tried putting the gpu in a different pcie slot on the motherboard but there was no change.

Im not sure what I can do about the possibility of failing capacitors/memory chips. Can this sort of thing be repaired or will i need a replacement?

Also if i were to take my graphics card and try it in a different PC would this remove any possibility of the PSU being at fault?