Striped(/black/blue/red) screen of death

Otthon Gamelj

Reputable
Jan 3, 2015
26
0
4,530
Greetings everyone who tries to help me!

I have a homebuilt PC for 1,5 years.
I bought a second-hand GPU from a previous CrossFire chain, a Sapphire 7950 with boost, Dual-X edition. Its fans are not the genuine ones, so it runs at 80-85 Celsius degrees.

Before the 7950 I had a 7850 from Sapphire, also Dual-X edition. It was perfect but I wanted something more powerful.

My PC specs are:
fx-8320
asrock 970 pro3 r2.0
4 gigs of kingmax ddr3-1600
the previously mentioned 7950
and windows 8.1 64 bit with its updates

everithing is going on their stock core clocks

I found out something. I underclocked the GPU to 850 MHz, because the power limit cannot be raised beccause it ran extremely hot that way (because of the budget fans on the gpu), so I left the power limit on 0% and underclocked the gpu. it has partly solved my problem so it didn't die this much on a day. but it still does this problem

Temps: cpu is cool, but the GPU goes on 80-85 Degrees.

It sometimes happens after 20 mins of gaming(FC4, Lichdom, AC:unity) but the day before yesterday I was able to game 5 hours of FC4 in a row.


My deaths look like:
link1

link 2
link3

These happens with 14.4 driver of Catalyst, but I had fully black screens too.

With the 14.12 Omega I had also these and fully black screen.


Memtest86 says everything's OK.
 
Solution
Those look like GPU memory artifacts. Your GPU memory is either borked or it doesn't have enough cooling. Whoever removed the stock cooler from the card fucked up and one or more of your memory modules on the card has problems. This doesn;t exclude the possibility of vrm problems because of the same issue.

You might want to take the card apart and reapply thermal paste and check the thermal pads on the memory and vrms.

Also these cards were used for mining virtual currency so chances are it was bios modded to draw less power. While it might have been stable mining its not that stable gaming. You might want to check BIOS version and if not original get an official bios (search google)

Also some cards have a physical bios switch (some...

emdea22

Distinguished
Those look like GPU memory artifacts. Your GPU memory is either borked or it doesn't have enough cooling. Whoever removed the stock cooler from the card fucked up and one or more of your memory modules on the card has problems. This doesn;t exclude the possibility of vrm problems because of the same issue.

You might want to take the card apart and reapply thermal paste and check the thermal pads on the memory and vrms.

Also these cards were used for mining virtual currency so chances are it was bios modded to draw less power. While it might have been stable mining its not that stable gaming. You might want to check BIOS version and if not original get an official bios (search google)

Also some cards have a physical bios switch (some well "hidden" ) on the card. Might want to try to switch that and see if it helps.

Good luck!
 
Solution

Otthon Gamelj

Reputable
Jan 3, 2015
26
0
4,530
Thanks for your answer. So you telling me that it's because of overheating? I'm planning to buy the stock fans for it, but they won't arrive in less than 3 weeks. I bought 8 grams of Arctic MX-02, it'll arrive on Monday, I'll use some of it for the GPU then.

I flashed a stock BIOS for it, it was one of my first movements.

It has a BIOS switch, actually, a huge one. And when it's in boost mode, it's lighting.
 

Otthon Gamelj

Reputable
Jan 3, 2015
26
0
4,530


Yes. The memory died or was dying on the card. I sent it back to seller and bought an XFX after it.