I was trying to use my gtx1060 in my rig which already have a rx580 8GB and a rx5700xt 8GB.
First, for some reason the gtx1060 raise an error
GPU driver error, no temps when I was using the “default” driver version which is
460.56, and it make the temp statuses for all GPU get mess(actually not being accountable). And also, it makes the AMD GPUs hashrate unstable.
Then, I’ve updated that to the most recent and stable, version
460.73.01, which cause me another and worse error:
CUDA error: an illegal memory access encontered (err_no=77), so for obvious reasons i did rollback my driver version change.
Then I’ve tried a most recent and not stable driver, which is
465.25 and Still getting the same Driver temperature issue, and since I am using the autofan, it make my GPUs got REALLY hot and then the system is rebutted for security threshold defined.
note: I tried a few different ways of configuration, the 3 GPUs in the same miner(nbminer), the AMD GPUs using TeamReadminer and the gtx1060 using the nbminer alone, etc…
My Question is: Do you guys know a stable driver version that I can use to not mess the things? Or even a configuration of miners to avoid this issue? I would not want to sell my 1060