Windows 10: GPU Driver constantly triggering TDR in various games.

I've rma'd my gpu twice (currently on my third). I've also replaced my PSU. Not sure how exactly to pinpoint what's been causing this shit. But basically over time (varying amounts of time per rma'd gpu) my gpu will more frequently trigger windows TDR. Eventually it'll start hard freezing my system with sound getting ****ed. Then it'll start freezing with all displays connected going a solid color (even my oculus rift which I thought was pretty funny). My first rma'd gpu eventually completely got borked and wouldn't post. Thankfully the gpu i'm on hasn't hit that point (seeing as my warranty is expired/was expired before EVGA decided to honor it to give me this one anyway). I originally thought it was a lack of power, so I rma'd my psu and nothing changed. Right now I'm thinking it's something in the mobo not supplying enough power through the pci-e slot (i've tried both slots to no avail) but I have no idea how to even begin to further diagnose this issue. I've previously run furmark with no crashes. I'll run P95 tonight. But provided are DM collector logs I ran shortly after a hard freeze preceded by three to four soft freezes within a single rocket league match.

P.S. Only one time did a hard freeze result in a BSOD and I'm pretty sure that was on a previous gpu so I doubt it'll be in the logs.

:)

 

display driver TDR issue - game(s) do(es)nt recover since windows 10

i have been experiencing display driver TDR errors for a long time, most likely due to my old notebook's bad cooling (HP envy 17 3D 1st gen - i7-720qm + mobility radeon 5850), as the drivers are being kept up-to-date all the time and my other hardware is
perfectly fine and no OC is done.

though by repasting cpu+gpu about 2-3x a year and cleaning the inside monthly, i have reduced these issues to a minimum, not happening for several weeks or months at all.

judging from the temperatures, it is not even the gpu which causes the TDR of the graphics driver, but the cpu, since it gets much hotter than the gpu.

however, the temperatures NEVER cause a heat-related restart, ie. they never reach the hardware specified Tjmax. the cpu just throttles down a bit, but this seems to cause the TDR error with the graphics driver when both, cpu+gpu are stressed, which only
happens during gaming (and there even not with every game).

but since i have upgraded from win 7 ultimate 64bit to win 10 pro 64 bit, when the "display driver has stopped responding and has recovered" error occurs, upon driver reset the game (im mostly playing cs:go and a bit diablo 3 where it happens) doesnt recover
anymore. in win 7, it used to freeze for a second, then the display would turn black and the sound would start to loop, and after a few more seconds the game would continue.

since windows 10, the game just freezes and although the driver recovers (alt-tabbing resp. ctrl-alt-del-ing shows the desktop), the game doesnt recover anymore, hence i need to terminate it. this is really annoying, since just a short freeze is bearable,
but having to terminate the game each time it happens is just bad.

changing the TDR-related registry values did not help at all.

so is there a way to make the game recover upon driver recovery as well, as it has been the case in win 7?

thanks

 

Nvidia drivers suddenly start to crash constantly

"Could it be an energy issue? I think the day of the problem the energy service was delivering poor voltage. "

Yes, inconsistent mains power can cause TDR\BSOD. So can cheap power supplies.

"I set the tdr to 8 seconds. Can that affect the performance? "

8 seconds is good, and the timing before TDR does not impact on performance.

Longer time before TDR could risk damage to the gpu if a TDR is actually caused by a problem with the gpu, such as

overheating though.

TDR is pretty much a failed system of preventing premature BSOD caused by lag, which only rarely works as intended,

and it always puts the blame on the drivers, but the lag is more often caused by software conflicts or other hardware issues

such as you found when disabling turbo-boost, which also may be symptomatic of your unstable power.

I've only had a handful of TDR errors and they were all on Win7. The only one I remember was with Crysis 2 which had a couple of

random TDR when using graphics settings that pushed my gpu to 100% use. When I lowered the games resolution a notch,

which lowered gpu use to 98 %, there were no more TDR errors .

Generally gpu's should be able to run at 100% use, but this game may have been an exception. However, thinking back on that,

the real issue may have been pushing the VRAM too far rather than the gpu usage (would have been using a GTX560ti 1GB then) ?

- Good old SR3... A great, stupidly violent and PC incorrect game... Loved it.

btw - It wouldn't affect your 1070, but this game is a good example of how a game can have one setting that can heavily affect

performance. It has a lighting setting that hit my frame rates heavily if I set it above low (or was it medium ?).

I could run the game with all other settings on max, but this one must have been poorly optimised.

I would have been using a GTX 770 2GB at the time.

- Looking at my files I see I used a few mods to fix a couple of aggravations - 'Just Get In The Damn Car', 'HUD fix', 'Advanced Launcher',

and 'Disable Tutorials'

 

display driver TDR issue - game(s) do(es)nt recover since windows 10

hi, thank you, but judging from your suggestions, you dont seem to have read what i have already tried...

"chipset drivers are up-to-date as well (intel chipset software installation utility 10.1.1.14)." - 10.1.1.14 is from
2/4/2016, ie. up-to-date.

and from my initial post...

"changing the TDR-related registry values did not help at all." - this implies changing the TDR delay value, as well as some other (TDR level etc.)

so i have tried all these already.

and i am talking about the game not recovering after the display driver has successfully recovered, not the driver itself, since the driver recovers perfectly each time it suffers a TDR error.