RTX 2080ti clocked high and crashing games

[KingAngelo]

Recently I have been crashing when trying to play games. I can’t even play 10 min before I get D3D device lost error. Any advice or help would be greatly appreciated!

This is a screenshot of my clock speeds when idle on desktop screen

 

[Waldorf]

any driver/win updates before this happened?
using any ocing, man/auto?
what psu? age?

make sure to fill out hw specs on your profile, make it visible..

 

[KingAngelo]

any driver/win updates before this happened?
using any ocing, man/auto?
what psu? age?

make sure to fill out hw specs on your profile, make it visible..

Windows has updated and all my drivers are up to date.
Havent OC my gpu but I believe my CPU came OC
I have a 750W PSU and ive had it from around 2018 I think

 

[Waldorf]

@KingAngelo
i overlooked it: turn off rebar in bios.
and please fill out the system page on your profile, easier for ppl to see.
(i can barely read your rig info pic, unless i crank up brightness, what you get when ur old, lol)

not so much if up to date, but when those updates (win/drivers) were applied,
as sometimes that can be an issue, so its good to know if it was working prior (any) update or not.
i recommend turning of any driver update from win/Nv panel, to make sure you have better control.

removing any oc, not talking about "factory" oc like higher (boost) clocks on stuff,
is a good idea for trouble shooting.

more important, what brand/model is the psu.
its getting old now, and while it might work, there will be a huge difference between
and old "tier 1" unit, and something like a 80 white certified "no-name".

i had a lnew ower tier 850w for testing, that would trigger my gpus on board diagnostic leds (power),
while a good 2y old 650w would not, even when stressed with cpu using max values (for the chip) PBO,
and the gpu oc with maxed power/voltage.

download the test tool
OCCT

start with the 3d std and enable error detection.
try vram test after that, then the power one if all above dont crash system.

make sure to end/disable all tools you use for ANY monitoring/tweaking/ocing, especially if running in the background.

 

[Beginner Macro Device]

its getting old now

5 years is nothing for a PSU. If it starts ageing at this point it's a very bad quality PSU, or it's seriously abused (non-stop mining at 90+ % power for example). I still have a bunch of 15 to 20 y.o. PSUs running smoothly, with only one of them requiring a one capacitor replacement one year ago. It's still totally capable.

I'd start with disabling all OC and installing the fresh OS on a separate SSD. The problem persists, you need to fix/replace your GPU. The problem goes away, you need to reconsider your overclocking and/or OS integrity.

 

[Waldorf]

and there are folks with units "going bad" with less than that, just because it doesnt happen to all, doesnt mean it wont ever.

constantly offering "fresh install" isnt really helping.
UNLESS everything else was done, which wasnt (yet).

enough times were i worked on mine/friends pc, where i was close to do it,
but decided to uninstall chipset/gpu/audio drivers, clean with ddu, and do fresh install with newer (or older known working) drivers,
only to find the issue fixed.

you seem to assume ppl only install win/drivers and a single game.

for me (ignoring i have everything split on six M2), if i had everything on one drive,
the 8 games that are "online only" would require to DL almost 400GB after fresh install.
so, yeah, great advice..

 

[phanbuey]

Look at the event manager to see what's going on. Try stress testing the card

start with the 3d std and enable error detection.
try vram test after that, then the power one if all above dont crash system.

^ this. D3D device lost is definitely due to the card crashing and the driver resetting, I usually see it when OC'ing ram too high.

Unfortunately ram chips on the 2080tis are notoriously fragile.