• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Progressive death of GPU?

S

Supertaquito

n00b
Joined
May 4, 2014
Messages
11
So this is what occurred in the last unfortunately days:

My sisters fiancee asked me to build him a computer to play League of Legends and the Elder Scrolls Online without any flinch. So I decided to build him a compact i5 3470 system with a Matx asus board with an h87 chip, 8gb of kingston ram, a cooler master silent pro m2 720 watt PSU and a to be R9 280X gpu.

He went looking for the parts and couldn't find a 3470 chip, so he decided to shed the extra bucks and bought an i7 3770 chip instead. I proceeded with building the computer and installing the OS. Later I decided to use my own XFX R9 280X GPU to give him an idea of what the end result would be in terms of video performance. I installed skyrim and modded it to the same extent my own skyrim is modded. This includes ENB, Realistic Lightning Overhaul and other 186 mods.

I did forget to install MSI afterburner to control my fan speeds. And started skyrim, during the opening scene, the screen suffered corruption and later a bluescreen, the case was very hot and the heat was coming from the GPU. I restarted the computer and set up the correct fan profile. No other issues were encountered from that point.

After he was happy with his to be computer, I returned my GPU to my own system which is a (i hope you guys do not cringe)

A10-5800K system with a gigabyte a85x motherboard.

While there is some bottleneck, my system handles such a massively skyrim instance incredibly well at 50fps average, I am guessing thanks to enboost and the GPU's power.

Anywho, I spent my weekend playing skyrim as I didn't have any plans with the SWMBO, on monday, new drivers came out which I installed (14.4) shortly after, whenever I would play skyrim, the same issue that occurred with my sister's fiance's computer started occurring with mine, with the exception that there were no temperature issues this time. The card would simply corrupt my screens, and not go into a BSOD, only leave my system useless and forced to restart. This happened with this driver, and the one used before which was the previous official WHQL. Upon starting up with the previous drivers, I noticed my skyrim's performance fell hard from 50 avg to 28 avg, to my surprise I noticed my GPU downclocked itself from a 1000 core and 1500 memory settings to 500 core and 750 memory on its own.

I reset its settings and proceeded to play skyrim, again, corruption. At first I thought that skyrim did something, so I tested Tomb Raider 2013. It worked for longer, then bam. Corruption again. The issue progressed to having hard flicker even when playing HD youtube videos. I decided to take the GPU out, leave the same drivers, and opt to use the inboard graphics.

No more issues were had. I have now filed a support with XFX and am a little scared they will deny the RMA because it started with an overheating issue that I have heard is not covered by the warranty.

Anywho... the question (yay, finally) does this look like an awful coincidence between the installation of drivers and the physical damage of my GPU? It seems to me that as no more issues were had with the same drivers after the card was taken out, it points towards a damaged GPU.

Or perhaps de GPU was too happy to be in an i7 system to be utilized properly and killed itself after it went back to my puny system.

Anywho.. any information which helps me find closure on my GPU's demise would be greatly appreciated.. as well as a little inside on XFX' RMA process (outside the U.S., I am in Mexico)

Thanks!
 
Someone needs to conduct a study to determine if GPU's can suffer depression.
 
Ultima99 said:
Someone needs to conduct a study to determine if GPU's can suffer depression.
Click to expand...
Yep, they need a quality elecstasy supply :p


Op, use GPU-z to check temps of vrms as well.
Use a finger on the back of the card to see if any of the memory chips are getting hot or if any other place is very hot.

What temp is the GPU reaching when the corruption starts?
What does the on screen corruption look like.
Is it in square blocks or lines across the screen?
 
mi7chy said:
Was it overclocked at any time?
Click to expand...
It was overclocked only once for a test when I first got it (back in Novemeber '13), but I didn't get any improvement on performance so I set it to stock. The overclock was 1020 and 1520 for core and memory respectively.

Other than the single day, no.
 
Nenu said:
Yep, they need a quality elecstasy supply :p


Op, use GPU-z to check temps of vrms as well.
Use a finger on the back of the card to see if any of the memory chips are getting hot or if any other place is very hot.

What temp is the GPU reaching when the corruption starts?
What does the on screen corruption look like.
Is it in square blocks or lines across the screen?
Click to expand...

You know, you got a point. I never did see VRM temps during testing, but the card will start failing under the following circumstances:

Temperature: 70C
GPU use: 85%

The corruption can be two ways, the screen will take the color of the most predominant texture on the screen at the time, and with gray, black or a lighter color of the background corruption vertical lines.. example:

https://lh5.googleusercontent.com/-.../AAAAAAAAAnE/NATdPHfljA0/w1216-h912-no/14+-+1

https://lh6.googleusercontent.com/-.../AAAAAAAAAnA/bKjMQZfES40/w1171-h878-no/14+-+1

https://lh6.googleusercontent.com/-.../AAAAAAAAAns/pJizm0x93Fc/w1171-h878-no/14+-+1

It can also present itself as black ''checkers'' on some section of the screen where the game produces lightning.

At the time I cannot test the card anymore as it has been packed to be returned if XFX accepts the warranty. I took hundreds of close up pictures of the PCB in case XFX claims it was damaged.
 
Your memory is failing. (gfx card memory)
Corruption in the shape of patterns of blocks and/or blocks of different colours are caused by the colour overlay being corrupt.
The colour overlay only exists in that format (that you can see) when on screen and as such is the memory failing because the screen is a window on the memory.

Memory failure can be caused by heat, faulty chip(s) or improper voltage (and some other things beyond your control).
Check your PSUs 12V lines with a multimeter to make sure they are well within 12V +/- 5% and that they dont drop much when under load.
 
Last edited:
Nenu said:
Your memory is failing.
Corruption in the shape of patterns of blocks and/or blocks of different colours are caused by the colour overlay being corrupt.
The colour overlay only exists in that format (that you can see) when on screen and as such is the memory failing because the screen is a window on the memory.

Memory failure can be caused by heat, faulty chip(s) or improper voltage (and some other things beyond your control).
Check your PSUs 12V lines with a multimeter to make sure they are well within 12V +/- 5% and that they dont drop much when under load.
Click to expand...

Thank you for your recommendation. It makes me think the PSU on the computer I built may have delivered improper voltage to it as the issues started shortly after being on that new computer.

I did forget to mention the card makes a whirr noise before the corruption occurs. It may be an electrical whirr, but it is not coil whine.

Do you have idea where I could find a multimeter?
 
Supertaquito said:
So this is what occurred in the last unfortunately days:

My sisters fiancee asked me to build him a computer to play League of Legends and the Elder Scrolls Online without any flinch. So I decided to build him a compact i5 3470 system with a Matx asus board with an h87 chip, 8gb of kingston ram, a cooler master silent pro m2 720 watt PSU and a to be R9 280X gpu.

He went looking for the parts and couldn't find a 3470 chip, so he decided to shed the extra bucks and bought an i7 3770 chip instead. I proceeded with building the computer and installing the OS. Later I decided to use my own XFX R9 280X GPU to give him an idea of what the end result would be in terms of video performance. I installed skyrim and modded it to the same extent my own skyrim is modded. This includes ENB, Realistic Lightning Overhaul and other 186 mods.

I did forget to install MSI afterburner to control my fan speeds. And started skyrim, during the opening scene, the screen suffered corruption and later a bluescreen, the case was very hot and the heat was coming from the GPU. I restarted the computer and set up the correct fan profile. No other issues were encountered from that point.

After he was happy with his to be computer, I returned my GPU to my own system which is a (i hope you guys do not cringe)

A10-5800K system with a gigabyte a85x motherboard.

While there is some bottleneck, my system handles such a massively skyrim instance incredibly well at 50fps average, I am guessing thanks to enboost and the GPU's power.

Anywho, I spent my weekend playing skyrim as I didn't have any plans with the SWMBO, on monday, new drivers came out which I installed (14.4) shortly after, whenever I would play skyrim, the same issue that occurred with my sister's fiance's computer started occurring with mine, with the exception that there were no temperature issues this time. The card would simply corrupt my screens, and not go into a BSOD, only leave my system useless and forced to restart. This happened with this driver, and the one used before which was the previous official WHQL. Upon starting up with the previous drivers, I noticed my skyrim's performance fell hard from 50 avg to 28 avg, to my surprise I noticed my GPU downclocked itself from a 1000 core and 1500 memory settings to 500 core and 750 memory on its own.

I reset its settings and proceeded to play skyrim, again, corruption. At first I thought that skyrim did something, so I tested Tomb Raider 2013. It worked for longer, then bam. Corruption again. The issue progressed to having hard flicker even when playing HD youtube videos. I decided to take the GPU out, leave the same drivers, and opt to use the inboard graphics.

No more issues were had. I have now filed a support with XFX and am a little scared they will deny the RMA because it started with an overheating issue that I have heard is not covered by the warranty.

Anywho... the question (yay, finally) does this look like an awful coincidence between the installation of drivers and the physical damage of my GPU? It seems to me that as no more issues were had with the same drivers after the card was taken out, it points towards a damaged GPU.

Or perhaps de GPU was too happy to be in an i7 system to be utilized properly and killed itself after it went back to my puny system.

Anywho.. any information which helps me find closure on my GPU's demise would be greatly appreciated.. as well as a little inside on XFX' RMA process (outside the U.S., I am in Mexico)

Thanks!
Click to expand...

If you haven't received your RMA yet, please PM Me.

Thank you
 
Are you grounding yourself when handling hardware? Do you do any of this on a carpeted floor or with socks?

It's often overlooked, but ESD is real and can damage hardware.
 
Talonz said:
Are you grounding yourself when handling hardware? Do you do any of this on a carpeted floor or with socks?

It's often overlooked, but ESD is real and can damage hardware.
Click to expand...

Yes, I ground myself by touching the chassis or a plugged PSU when working on hardware.
 
You must log in or register to reply here.
Back
Top