Cuda 10 I'm using, as far as I know cuda 11 is only for newer NVidia cards.... and it's not working for 1080ti. However I have 10 rigs with 6x 1080ti, only one of them is making this kind of issues.
Config: -a ethash -o stratum2+tcp://daggerhashimoto.eu.nicehash.com:3353 -u [wallet] -w $rigName --kernel 0 --mt 6
OC: core +100 / +125 depends of the card;
memory +600
PL 210W
and the OhGodAnETHlargementPill activated.
It's not 90 Watts, this is the SW readings.... I get around 1400w at wall for the whole rig (6x1080ti) so I would say 220W for one 1080ti.
What driver version? I don't think You need OhGod and --mt option , they basically do the same.
I believe that using Nicehash is more stressful for Your rigs because of more often DAG change and creation. Try pointing problematic rig to Eth pool to see how it behaves.
Driver version 430.40 on linux. You're right that mining to an ETH pool is more stable and as far as I remember never has this problem mining on ETH pools... however, my question here is why the rig is not restarting since the Watchdog was triggered?
If it's restarting, then i have no problem, loosing only some seconds to mine, but if it hangs up then it burns electricity for 3-4-5 hours maybe....