Maybe someone ran into this before or knows whats going on.

I have a 980ti KPE, where the instant the GPU Temperature hits -64C the Driver crashes. What i find interesting is that at that temp the monitoring also starts saying " NOT GOOD".

I have tried this with the with the stock oc bios, the stock ln2 bios and the XOC bios, all same result. I found nothing in the public guide about this.

 

Appreciate any help!

isn't this just an issue with new drivers? If use old enough drivers the CB goes away AFAIK

Yeah the driver issue was considered, but on all other maxwell cards i have:

- the crash is different

- the temperature has some range and is not exact

- there is a tweak which lowers the temperature where it happens and that made 0 difference on this card

Not to mention the "NOT GOOD" would have been programmed before cuda 10 came out.

If noone has an idea i will retest with old driver but i think its unlikely.

Update:

After talking with Tin and doing some more testing, it was indeed the new driver causing the coldbug.

The "NOT GOOD" message is cause by the tmp411 reaching its minimum temperature and just a coincidence that it happens at the same temp.

What i still dont understand is why the behaviour of the coldbug is so different from 970/980 non ti. Looking at the ranking it looks like its not just my card .