So far, I have 2 units with errors. They will be soon the too many error units. I have 3 units completed successfully, and one unit that soon will complete.

These units are causing more frequent down clocking (several times in the last day as oppose to once every few weeks) on my windows 10 machine and it happens on card 0 (I have 2 980 ti cards on this machine), which happens to be the card connected to my monitor.

Like I said before, this doesn't happen in such frequency, when running other units. I would conclude that these units more sensitive than most, do need some refining.

So far, I have 2 units with errors. They will be soon the too many error units. I have 3 units completed successfully, and one unit that soon will complete.

These units are causing more frequent down clocking (several times in the last day as oppose to once every few weeks) on my windows 10 machine and it happens on card 0 (I have 2 980 ti cards on this machine), which happens to be the card connected to my monitor.

Like I said before, this doesn't happen in such frequency, when running other units. I would conclude that these units more sensitive than most, do need some refining.

Since I upgraded to driver 381.65, I have no more down clocking problems, so far. It's nice to get one benefit out of this.

Since I upgraded to driver 381.65, I have no more down clocking problems, so far. It's nice to get one benefit out of this.

for me, just the opposite is the case: after updating my two Windows 10 machines to driver 381.65 last weekend, in order to make it work with the latest crunching software (CUDA80), the GTX970 as well as the GTX750ti seem even more susceptible to overclocking.
Interestingly enough, it's different between the types of WUs; but in general, as I said, the GPU clock falls back to the default value (1152MHz for the GTX970 and 1058MHz for the GTX750ti) very quickly.
Then, no change of the clock is possible until the PC gets rebooted :-(
Quite a tedious situation.

This problem I never ever had with XP so far.