I'd start by dialling back the overclocking on all cards to 1080 and see if they all hold at that, then one by one increase them to see if they hold 1100. Personally I prefer running all cards a little below the max, mixed opinions on this but for 24/7 running I don't think it's wise to push the cards especially reference Vegas with the pissy drivers and high temps.
For me card temps weren't related to the GPU COMPUTE ERROR, as long as the cards were clocked stable high temps only reduced mining speed in my experience.
Hey everyone, i am new to this forum. Out of 6, one of my vega 64, when installed BC driver and after running PP registry, went full throttle on fan speed, and windows error shows up showing gpu temp overload and my system hangs. Its just with one card, no matter what i do, on normal driver it works fine, but on BC driver its creating this problem. Is my GPU faulty, should i return it ?. Please help
Start by dialling back the overclock/undervolt in that card. Not all cards can hold the same speeds/voltages. One of mine always runs 10% slower than the rest. I change the tach lights to blue on the weaker card/s that way when new drivers are installed or different mining software used and the order gets messed up you can track the weaker card easily