GTX 980 dying? Multiple symptoms

[xsater]

Hello, there

 

I think that my GPU might be dying..There are multiple different symptoms, which might be caused by different things, but the recent one made me think that it is probably the GPU.

Full specs:

CPU: i7 4790k (not overclocked at the moment)

MB: ASUS Z97 Sabertooth Mark S

GPU: Galax GTX980 HOF (it was overclocked for a very short period of time)

RAM: Kingston DDR3 HyperX Fury 1866MHz (4 sticks x 4 Gb)

PSU: Thermaltake Toughpower XT Platinum 1275W Snow Edition

Storage: SK hynix SH910 - OS, Samsung 850 EVO - games, WD Black - storage

Dual monitor setup. Both connected by DP.

 

Symptoms:

1) A few month ago, my PC started to shutdown randomly, while gaming. This was caused by ASUS anti-surge protection. I've done some googling and turned it off in UEFI. Might not be the smartest thing to do...

2) After that it was OK for a few month. And then...Sudden PC freezes while gaming. Complete system freeze with a crazy sound loop, keyboard and mouse are not functioning. Sometimes even the reset button is not working, so I have to hold the power button to shut down. The game itself is not demanding. It is an ~8 years old MMORPG. This happens more often that the next 2 symptoms;

3) Nvidia drivers stopped responding...First both monitors go black, then a couple of seconds later everything is back to normal. Sometimes it does not go back to normal. Primary monitor which has a game on is still black, while the secondary one is OK;

4) Pixelation thing (artifacts?) that can be seen on a screenshot below. It only happens on a primary monitor, which displays a game. I can still play the game, restarting the game brings it back to normal.

Spoiler

 

What I've tried so far:

1) Different driver versions - 385, 353, 362. I've been using DDU to remove the drivers before installing another version. 

2) Remove braided PSU extensions and connect GPU power directly using original PSU cables. While doing so, I found out that 2 pins on an 8-pin connector were 2-3mm off. For better understanding, I've made a reconstrucion of how it was and took a photo:

Spoiler

I plugged everything properly and this fixed a problem for a whole day. Then it came back again =)

Also, I noticed that there was a single power cable for 2x8-pin power connectors. So I decided to try and use 2 separate 8-pin cables;

3) Swapping GPU. So I took GTX960 from another PC and it worked for 1-2 hours without any problems. Probably will test it for a couple of days next time, I shall also put my GPU in another PC and see what happens there;

4) Increasing GPU voltage in MSI afterburner. Did not help;

5) Decreasing GPU clock in MSI afterburner. Did help for a few hours;

6) Cleaned the GPU rad and fans, re-applying the thermal compound.

 

So...Is it really dying? =(

And how could I know for sure? 

 

 

[Boomwebsearch]

3 minutes ago, xsater said:

So...Is it really dying? =(

And how could I know for sure? 

I would not worry too much about GPU failure as it is not as bad as if other components were to fail and cause damage such as if a hard drive were to fail data loss could occur. You could install HWmonitor and see if it can detect a failure in the component of your graphics card or potential failure. 

[Projex]

Did you overclock the card? If you did, that is the cause of this. Put down the overclock a bit, and you will be good.

[xsater]

5 minutes ago, Boomwebsearch said:

I would not worry too much about GPU failure as it is not as bad as if other components were to fail and cause damage such as if a hard drive were to fail data loss could occur. You could install HWmonitor and see if it can detect a failure in the component of your graphics card or potential failure. 

I actually do have HWmonitor installed. But what should I be looking for there? When my PC freezes completely, screens might go black, so I am unable to see the latest stats.

 

5 minutes ago, Projex said:

Did you overclock the card? If you did, that is the cause of this. Put down the overclock a bit, and you will be good.

I've been using the system for almost 2 years now. During this period of time the GPU was only overclocked for a month or so. Right now I actually downclocked the card, hoping that it might help. Which did..But not for long =)

[Boomwebsearch]

You can install FurMark using a link which I will provide and make sure that your GPU is delivering the performance that it should be delivering.

 

download FurMark using the following link:   http://ozone3d.net/benchmarks/fur/

 

Compare the results you get to the score of 6724 (what your GTX 980 should score). If you get a result which is lower then that is a good indicator that your card is going to fail within an unknown span of time.

[xsater]

I will try that out. Thanks!

[Qi_Forever]

Interesting, I often wonder how power outages and hard resets effect computer components.  Since I had a expensive CPU just die on me for no apparent reason, last year. 

 

So the BIOS gave an anti-surge warning?  or windows gave anti-surge warning?

 

Could be the power supply failing to provide enough power to the GPU, after an incident like that. 

 

Can you swap the GPU to another system to test it or can you swap out the PSU and test it?

 

I'm also in the opinion Nvidia hamper the drivers for ageing cards, this is pure tin foil hat theory though.  (will it be the Apple slowing phones down scandle in next few years, watch this space) 

 

[xsater]

2 minutes ago, Qi_Forever said:

So the BIOS gave an anti-surge warning?  or windows gave anti-surge warning?

The BIOS. Actually I forgot to mention. I've also updated the BIOS trying to solve the issue. 

After updating it, the anti-surge turned back on and I got the anti-surge shutdown once again. 

 

2 minutes ago, Qi_Forever said:

Can you swap the GPU to another system to test it or can you swap out the PSU and test it?

Yes, I will try that as soon as I get back from holidays

 

27 minutes ago, Qi_Forever said:

I'm also in the opinion Nvidia hamper the drivers for ageing cards, this is pure tin foil hat theory though.  (will it be the Apple slowing phones down scandle in next few years, watch this space) 

This theory actually makes a lot of sense.. =) But I wouldn't want to start this kind of discussion here

[xsater]

On 02.01.2018 at 11:11 PM, Qi_Forever said:

Can you swap the GPU to another system to test it or can you swap out the PSU and test it?

So..an update..

Swapped GPUs with my GF..

Now I have Gigabyte GTX 960 Windforce OC in my system.

 

Her PC specs:

CPU: i5 4690

MB: MSI H97 PC Mate

GPU: My Galax GTX980 HOF

RAM: Kingston DDR3 HyperX Fury (2 sticks x 4 Gb)

PSU: Cooler Master Thunder M 520W

Storage: SSD + HDD

Single monitor setup connected by HDMI

 

We've been playing the same game for 2 weeks now. So far she had that pixelation thing that I've mentioned in the 1st post, but only twice (during this 2 week period of time). Other sympthoms were not present on her PC. Average play time was ~3 hours per day. Compared to when I had GTX980 in my system, she has no problems at all.

As for me - I had 0 issues with GTX960 during these 2 weeks.

 

This is getting interesting...

[Ithanul]

Have to consider, a GTX960 does not pull near the same amount as a GTX980.

Plus, the fact the anti-surge keeps going off makes me think the PSU is going bad.

 

Sees Thermaktake, yeah, I would try to see if you can get your hands on another PSU to test with.

How old is that PSU?

[xsater]

32 minutes ago, Ithanul said:

Have to consider, a GTX960 does not pull near the same amount as a GTX980.

Plus, the fact the anti-surge keeps going off makes me think the PSU is going bad.

 

Sees Thermaktake, yeah, I would try to see if you can get your hands on another PSU to test with.

How old is that PSU?

Actually, I was thinking in this direction..

I've looked through the specs of GTX980 and GTX960 a week ago to find out the amount of amps required for the GPU..

My PSU is almost 2 years old..btw what's wrong with TT PSUs? 

I will try to get another PSU for testing..let's hope for the best

 

 

[Ithanul]

24 minutes ago, xsater said:

Actually, I was thinking in this direction..

I've looked through the specs of GTX980 and GTX960 a week ago to find out the amount of amps required for the GPU..

My PSU is almost 2 years old..btw what's wrong with TT PSUs? 

I will try to get another PSU for testing..let's hope for the best

 

 

I only had experience with one TT PSU which the model was a 850W Black widow.  Ran good for two years then made a loud bang like a shotgun and died.

Out of all my PSUs, it the only one to die from use.  I am unsure of the OEM for the model you have, so the internals may be better or not.