Possible GPU failure?

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.

StephenSchneider95

Honorable
Mar 14, 2016
52
0
10,530
This problem just started. The computer will work fine when doing any internet browsing to include streaming video and music. When I load a game the screens enter power saving mode, I can still hear all in game sounds as if it loaded properly, but I am unable to get the monitors to "wake up" without restarting the system. I have tried a few different games to see if maybe it was just high intensity graphics that were causing it. It appears to be all games at this time.

Any ideas are welcome! Thank you!

Edit: Games tested with same result: Minecraft, Fortnite, League of Legends, and Black Desert Online.

After replacing the GPU fan games ran for a fair amount of time, after about 3 hours of gaming got black screened again, since then its within minutes of opening any game. So I'm back to square one.

I ran a FurMark test to see if I could identify a reason for the screens blacking out. The temperature of the card stays below 66ºC, so I don't think temperature is an issue. The only thing that seems common before the screens black out is the load hits 98%-100% then a few seconds later it happens. The screens did go to power saving mode while running the FurMark test and were unable to be "woke up". (https://imgur.com/a/LB3f6).

I also ran a Heaven Benchmark 4.0, PC made it 175 seconds into the benchmark test running 99-100% load staying 66ºC before monitors entered power saving mode.
 


I'm no computer guy, but I'm assuming that when the GPU load was increased and the fan speed was required to increase it was stalling out or just not increasing enough and at that point the GPU was giving up.
 


My current PSU: EVGA SuperNOVA 1000 P2 220-P2-1000-XR 80+ PLATINUM 1000W Fully Modular EVGA ECO Mode. As for the GPU connection, it has two 6pin ports (https://imgur.com/a/9fqi6). I have two separate cables that go into this, both are 8pin cables with 2pin pull away. I should mention that I've had this card for going on 3 years. We did dust it really well, took the shroud off and cleaned it thoroughly, but no we didn't do anything with thermal pads. Temperature seemed fine on MSI Afterburner, could this still be the issue?
 


Thats not true. I have seen GPU failure only affect its ability to do 3d accelerated apps/games.
 


What would you recommend to verify if that is the case?
 
Well i've left once my R7 250 without heatsink and it just cutted off display, didnt check for sound but i know for sure it was responsive cause of numlock.
I think its a shot to change thermal pads, if the GPU doesnt have sensors there (as it doesnt have, only better brands do have like msi and gigabyte), my tought is try it.
 


Install Afterburner. Reduce the GPU clock speed by 100mhz and apply. Then try again.
 


I will try this this afternoon, still early here so no stores are open to try and obtain thermal pads.
 


Lowered GPU clock speed by 100mhz and ran FurMark again. 12 seconds in monitors enter power saving mode.
 
In power options did you set high perfomance and checked under advanced tab that everything power feature is OFF/DISABLED?
After you get to power options, select "Change plan settings" then -> "Change advanced power settings".
Now under Sleep select never everything (just to be sure), and PCI Express->Link State Power (OFF).
And last tab When sharing-> Prevent and also When playing -> Optimize.

Did you try to re-plug monitor? Or to turn off then turn on? DId you try to alt+tab when that happens?
Did you try to swap monitor cable or try TV?
Did you try lower resolution (on desktop)?
Tried disabling nvidia Geforce [strike](that thingy for recording, currently i cannot guess the name cause it gives me error of update...[/strike] IN-GAME OVERLAY)?
 


Followed the first portion of your instruction and everything is set correctly as you stated.
I have tried re-plugging the monitor, no change. Same with turn off then on. No result with alt+tab.
I have swapped from a 1440 monitor to a 1080 monitor, the only difference was Skryim (the game I tested first) ran at a 77% load. So I ran a FurMark test, lasted 12 seconds at 99%-100% on the 1080 and entered power saving mode.
In-Game Overlay is now disabled.

Unfortunately none of these changed much as it still seems at a 98% and up load on the GPU is causing the issue. Whether it be from the PSU or the GPU itself.
 
There is another thing, well last that will i suggest you (soo i don't bother you), go to safe mode (Shift+restart) https://www.digitalcitizen.life/4-ways-boot-safe-mode-windows-10
Download DDU (clean+restart),
Now after you come back in windows, extract the drivers (not automatic, select your GPU and windows), then after extracting,
Device manager then display adapter (expand) then righ click in it (on whatever the name, generic, microsoft display adapter) update, then browse and again browse and find folder and click okay, and Next, it will take some time to install, after it says finished , just restart (not force, go to start menu). And try few games.
 


I think someone else suggested that early in this thread. I tried it but it didn't change anything. I'm completely stumped on the issue, unless it is something as simple as a PSU not kicking enough power to the GPU.
 
For anyone who stumbles on this looking because of similar issues. As of right now it is fixed and is surviving the benchmark tests. (https://imgur.com/KwrHXLb | https://imgur.com/0Mlh8yd). I moved one of the power cabled connecting my PSU to my GPU from VGA2 to VGA3 slot on the PSU. I didn't change the cable but i'm guessing its just a bad slot on the PSU.