Leo Fang
We are dropping support for CUDA 11 very soon. This is moot.
(Reminder to self: #275 & #279 cover CTK major/minor tests).
cc @pciolkosz @cryos for vis
Based on my learning from the CUTLASS team, we should be able to support this API easily (by passing `Kernel` and `LaunchConfig`): `cuOccupancyMaxActiveClusters`.
/ok to test c9fac0b
This is ready.
/ok to test d70ec240
/ok to test d279e50
Blocked by #687.