How much time does it take to switch to another thread? For example, consider the case where I begin some loading operation. For the processor to switch to another thread and execute some command, does that take 1 cycle?

@alexander I think that's going to vary a lot by machine and by how much data is associated with each thread but it's probably a lot more costly than a single cycle.

It says that additional threads lead to no benefit -- but what if something goes wrong on one of the first six threads? Does stalling always take exactly the same amount of time for a particular computation like is shown in this diagram? I would think it might be useful to have a few "backup" threads available -- this diagram looks quite tight