The difficulty level and dependency latency is specific to every architecture. Just about every SM subpartition and SM has other execution models such as load/keep units, double precision floating point models, half precision floating position models, branch models, and many others.While we are nevertheless waiting for more information about Navi,