Ahh I see, in this slide a horizontal scan tells us show us that most 2 colors will overlap at any given time which tells us this processor has only 2 ALUs.

@sagoyal, correct, except you meant a vertical scan. At any one time, only 2 of the 4 threads are colored.

If at one time at most #{fetch and decode units} threads are active, does it mean we can use at most this number of ALUs at one time? What if the number of fetch and decode units is less than the number of ALUs? Can we still make use of all ALUs?