Strange GPU crashing issue

[93theripper]

I've been having this issue for a while now, some games seem to crash every 10-30 mins, most to desktop, some (rarely) require a full reboot as the whole system crashes.

It's always similar sinthoms across different games. Most common is black screen while audio is still going. Some time it's a full application crash. When the game gives an error it's usually a gpu/driver not responding or a dx3d generic error.

The problem is solved in most of this situation by downclocking my GPU by -200\-200 (sometimes -200 on mem only is sufficient) i will provide a list to see if somebody has an idea on what the problem is. I run multiple gpu stress tests\benchmarks and gpu memory tests without issue and many game work perfectly even with the gpu overclocked.

Just for clarification the games crash on stock clocks.

 

Here is the list on what i have collected so far:

 

Baldurs Gate 3 - Crash to desktop, downclock fixes problem (could be because early access)

Ghostrunner Crash to desktop with dx12 error, downclock fixes problem

The Dungeon Of Naheulbeuk - The Amulet Of Chaos Crash to desktop or sys crash, downclock fixes problem. wierdly the game doesent utilize 100% gpu even with fps unlocked + crap performance until downclock

Wasteland 3 Crash to desktop or sys crash, downclock fixes problem. wierdly the game doesent utilize 100% gpu even with fps unlocked

Cyberpunk 2077 Crash to desktop, downclock on mem only fixes problem

Yakuza like a dragon Crash to desktop with some black "artefacting?" - no solution.

 

Intensive Games that work perfectly fine even with overclock: all UBISOFT games, DOOM, SotTR, Death Stranding, MortalShell. Also all VR games work perfectly (including very intensive ones).

2D indie games have no problems.

I'm going off of memory here and i may have missed some games but it's happening wierdly frequently. Is it a known dx12 bug or an engine bug? What's the link? Do i have a faulty gpu?

What i already tryed: DDU+clean driver install\rollback\update and clean boot (all overlays disabled) did not fix the problem. No overheating issues.

asus strix 2080ti, 7700k, 32gb ram, usually play at 4k with RTX if availeble.

I'm gonna swap my gpu for a 3090 shortly and i wanna be sure to not sell a faulty gpu to a friend.

Thanks for the help

[TylerDurden!]

20 minutes ago, 93theripper said:

games crash on stock clocks.

do you have vrm temps that show up in gpuz with that card?

[93theripper]

13 hours ago, TylerDurden! said:

do you have vrm temps that show up in gpuz with that card?

Yes, they seem to update only on application restart for some reason but it seems they cap at 77C on furmark and round the same ingame. Also see how the game behaves in the pic. only 61% gpu usage under 60fps, same behavior in 3 of the games above. It's strange tho that the VRM hits 80C while gpu is at 60ish %. But why only in some games, i don't understand.

 

UPDATE: Trying to stress vrm as much as i can, reached 85C seems to cap at that temp, hwinfo is saying performance limit - relaiability voltage "yes"

[Kupal]

16 minutes ago, 93theripper said:

Yes, they seem to update only on application restart for some reason but it seems they cap at 77C on furmark and round the same ingame. Also see how the game behaves in the pic. only 61% gpu usage under 60fps, same behavior in 3 of the games above. It's strange tho that the VRM hits 80C while gpu is at 60ish %. But why only in some games, i don't understand.

 

UPDATE: Trying to stress vrm as much as i can, reached 84C, hwinfo is saying performance limit - relaiability voltage "yes"

 

It because of game settings.. Temp GPU increase based on game demand / settings..

I Suggest clean your GPU and entire PC..

 

your 7700k CPU Temp alone on that crap game with 50% usage are already high.

 

[93theripper]

1 minute ago, Kupal said:

 

It because of game settings.. Temp GPU increase based on game demand / settings..

I Suggest clean your GPU and entire PC..

 

Ur getting both 76 C on ur CPU alone on that crap game.

 

I recently cleaned everything with compressed air, i also tryed with open panel, slightly better temps but doesent solve the problem. Why are some games not using full gpu and crashing baffles me.

[Kupal]

2 minutes ago, 93theripper said:

I recently cleaned everything with compressed air, i also tryed with open panel, slightly better temps but doesent solve the problem. Why are some games not using full gpu and crashing baffles me.

That not clean yet. I mean replace thermal paste. that CPU is too hot at that 50% usage.

 

To find out how to make sure GPU can be used 100%. Either Boot up to different OS.

[93theripper]

1 minute ago, Kupal said:

That not clean yet. I mean replace thermal paste. that CPU is too hot at that 50% usage.

 

i'm swapping to aio + 5900x in a week, gonna do that before i sell, still doesent explain the problem

[Kupal]

2 minutes ago, 93theripper said:

i'm swapping to aio + 5900x in a week, gonna do that before i sell, still doesent explain the problem

Its either Windows OS Problem (With Windows Update) or PSU Problem.

Everytime I encounter this kind of problem. I always have back custom win10 iso with window update/driver removed. So I just manually install Driver.

[93theripper]

Just now, Kupal said:

Its either Windows OS Problem (With Windows Update) or PSU Problem.

That could be, so you are saying it's not a gpu problem? That said, any idea on why only on certain games? the psu is a corsair 750w

[Kupal]

5 minutes ago, 93theripper said:

That could be, so you are saying it's not a gpu problem? That said, any idea on why only on certain games? the psu is a corsair 750w

It could be. But to make sure at least try to install a custom windows 10 (On Spare SSD Drive) with windows update/driver removed..You can do google search.

Or use Windows 7/Linux.. then see if the GPU usage is working properly on those OS. If its not then its PSU.

[93theripper]

20 hours ago, Kupal said:

It could be. But to make sure at least try to install a custom windows 10 (On Spare SSD Drive) with windows update/driver removed..You can do google search.

Or use Windows 7/Linux.. then see if the GPU usage is working properly on those OS. If its not then its PSU.

I tested on a fresh install, same issue

[Mark Kaine]

Sounds like faulty factory OC to me, might also be a vrm or other hw issue of course. 

 

 

 

[93theripper]

2 minutes ago, Mark Kaine said:

Sounds like faulty factory OC to me, might also be a vrm or other hw issue of course. 

 

 

 

how do you explain it only happening in some games?

[Mark Kaine]

13 minutes ago, 93theripper said:

how do you explain it only happening in some games?

because some games are more susceptible to (unstable) OCs than others (typically resulting in a *driver error *) 

[93theripper]

9 hours ago, Mark Kaine said:

because some games are more susceptible to (unstable) OCs than others (typically resulting in a *driver error *) 

i see, do you think i can rma it for this issue?

[Kupal]

On 1/3/2021 at 7:39 PM, 93theripper said:

i see, do you think i can rma it for this issue?

If it still in warranty. then RMA and tell em it has VRM Issue.