Hoping some of these logs/info help some of you out.
All cards running on Nvidia 388.59
All cards running on 0.5.7
Mining on Flypool US server
The 1080Ti and the 1060 are in my every day gaming rig with decent airflow. the 1080Ti has a back-plate attached.
The 1080 is in my wifes computer, decent airflow.
The 980Ti is on a test bench and kept in a fairly cool spare bedroom, the new 1080 will be going into my wifes computer and the old founders edition onto the testbench.

ASUS ROG STRIX GeForce GTX 1080 TI 11GB
+150GPU +500mem 120% power target 100% fan speed
61c AVG 769.9Sol/s 2.71Sol/w
Memory appears to help with the hash-rate up to +500, after that the gain is minimal.


EVGA GeForce GTX 1080 Founders Edition 8GB
+40GPU +300mem 90% power target 90% fan speed
64c AVG 524.6Sol/s 3.25Sol/w


ZOTAC GeForce GTX 1060 AMP Edition 6GB
+150GPU +500mem 120% power target 100% fan speed
54c AVG 329.7Sol/s 2.69Sol/w
Memory appears to have almost no effect by itself, however at +150GPU and no memory overclock it is unstable, it only becomes stable when I add in the memory overclock.


EVGA GeForce GTX 980 Ti 6GB FTW
+50GPU +400mem 115% power target 100% fan speed
58c AVG 495.4Sol/s 1.95Sol/w
The memory response appears the same as the 1060 whereas it doesn't seem to directly have an influence, but is needed for a stable GPU overclock. For some reason I can no longer adjust the memory overclock.

I have an EVGA GeForce GTX 1080 SC GAMING ACX 3.0, 8GB on order aswell.