[SOLVED] GPU needs to be downclocked to avoid crashing

[Scott Baban]

My GPU needs to be downclocked to a -50mhz on core clock and memory clock on Afterburner to avoid screen-freeze on games or just viewing videos.

The GPU is a ZOTAC GTX 1080 Ti AMP! Edition, bought it last 2 years ago. It was worked fine until last year. Is there a fix to this problem? RMA's out of the question as I've lost the receipt and the logistics would be too pricey for me.

System:
Ryzen 5 2600
B450M DS3H
16GB 3000mHz
GTX 1080 Ti AMP! Ed.

 

[Dj0gany]

On a 1080Ti,taking 50mhz from core clock and memory clock really wouldnt make a difference in gaming.
What are the temps?
Are there any articafacts before crash?

 

[Scott Baban]

On a 1080Ti,taking 50mhz from core clock and memory clock really wouldnt make a difference in gaming.
What are the temps?
Are there any articafacts before crash?

When playing DotA, I get an avg of 50-65 degrees celsius. No artifacts are found too, it just screen-freeze or crashes the game.

 

[Dj0gany]

When playing DotA, I get an avg of 50-65 degrees celsius. No artifacts are found too, it just screen-freeze or crashes the game.

Drivers all up to date?

 

[haseeb98ahm]

Factory OC are fully validated so it normally shouldn't crash, your temps are fine too...
What PSU are you using? Is it enough for your system? NVIDIA recommends 600W PSU.

 

[Scott Baban]

Drivers all up to date?

Yep current drivers are installed :/

 

[Karadjgne]

Normally having to downclock means either the cards VRM's or vram is failing. In this case since you have to downclock speeds as well as memory, I'd say it's the VRM's that are going. They may be fine, it might just be a case of them overheating at the factory OC settings, pushing too much amperage, but either way that's where I believe the issue is.

You could try replacing the heatpads, they aren't expensive, will also require repasting the gpu, but if that fails to fix the issue, then just like a hdd with a bad sector, it'll only be a matter of time before the VRM's fail completely and render the card unusable.

 

[Scott Baban]

Normally having to downclock means either the cards VRM's or vram is failing. In this case since you have to downclock speeds as well as memory, I'd say it's the VRM's that are going. They may be fine, it might just be a case of them overheating at the factory OC settings, pushing too much amperage, but either way that's where I believe the issue is.

You could try replacing the heatpads, they aren't expensive, will also require repasting the gpu, but if that fails to fix the issue, then just like a hdd with a bad sector, it'll only be a matter of time before the VRM's fail completely and render the card unusable.

I would try this soon, sadly I haven't found any computer tech near where I live. I don't trust myself replacing stuff that involves the internal parts. Thank you sir!