Random TDRs (display driver crashed and has recovered) with almost all games.

G

Guest

Guest
So, for a VERY long time I've been getting random TDRs and I've tried at LEAST 10 different drivers and I have also switched PCI express slots and still get the same issue. I sometimes get that problem with nonblizzard games but not as often (for example i played fallout 4 for 160 hours and haven't crashed except once), I tried all the basic troubleshooting steps as well like closing background apps, using different settings in nvidia control panel, running windowed mode, even reinstalled windows once. I heard from a few sources that games like sc2 and heroes of the storm have an issue with factory overclocked cards or something but in any case, i have no clue how to fix this and it has been REALLY depressing me lately so i would appreciate it if someone could help me out.
also something maybe worth noting is that it might be a problem with opengl perhaps? how do i test to see if it is?
In addition, I have an EVGA gtx 980 SC edition and I have seen several forum posts regarding TDRs with factor overclocked gtx 980s like this one
http://forums.evga.com/GTX-980-FTW-quotDisplay-driver-nvlddmkm-stopped-responding-and-has-successfully-recoveredquot-m2302108-p3.aspx
I've been getting this problem for SOOO long (almost a year now) and i'm still trying to figure it out. I could RMA the card, but I really don't want to do that since its a huge hassle for me and im afraid of getting another card with even more problems than this one.

error = Display driver nvlddmkm stopped responding and has successfully recovered. (screen suddenly turns into certain color like brown for a few seconds then the game comes back to life)

specs
EVGA gtx 980 SC
i7 3770k @4.2ghz
Corsair vengeance 8gb ram
Corsair TX850w psu
Msi z77agd65 motherboard
Samsung EVO 250gb SSD
 
use something like evga precision X to underclock the card to normal levels if the problem goes away you then have three options.

a) increase the voltage slightly
b) run it at a lower clock
c) RMA it

My preferred option would be C, especially if you can prove it is stable at non OC'd speeds.
 
G

Guest

Guest


how about using debug mode in nvidia control panel?
 
G

Guest

Guest
Evga customer support told me to try it, it sets the card to stock clocks apparently. It's under "help" in nvidia control panel. I'm also running the extended windows memory diagnostics tool now to check if it's the memory. Psu voltages are fine according to the bios also
 
G

Guest

Guest
Also I'm using msi afterburner + hardware monitor and I want to figure out if the gpu is taking up too much power when the crash happens or something, but idk how much power is too much power or too little.