AFAIK yes only for the whole workload. But that is value determined by testing live and not querying the OS. I'm not aware of each task storing any type of load. I'm not sure of any easy way to do it. I suppose you could run an interrupt on a timer and montor the current task to see what ones are mostly in use. The vertical blank was used for task switching IIRC. What would be good is checking at the same speed tasks are switched so you can catch the change. Or you could also run through the task list in an interrupt to see what ones are ready and waiting. I don't know if that would be more or less accurate but an interrupt shouldn't be used for any time consuming work or anything sensitive as causing a fault will freeze the system. Even though the task list would be fairly static during an interrupt.

You're welcome. But I am surprised. There has got to be other people on the site who have either programmed a CPU task monitor or know how it would be done. At least on AmigaOS. Another thing is to examine the task priority and enter this into the equation. Since tasks running on a higher priority than others will logically gain more CPU time. Also, there is Quantum and Elapsed fields of ExecBase. These contain how much time a task gets and how much it has used currently. IIRC this is controlled by a standard vertical blanking interrupt. So you could also keep watch of these values to see how much time a task is getting and if a task is hogging the CPU. Peeking ThisTask from an interupt will give you the current one. Each task wold get an even anount of time. AFAIK the quantum amount was 4 ticks. So if there were 50 ticks a second which is the standard for DOS BTW then each task gets an equal amount of 0.08 seconds each or 4/50. Or 12.5 tasks a second. If my math is correct. :-) Of course this is the time they have but not always the time they use with other factors involved such as a task waiting for signals and thus giving up ownership of their slice of quantums. And even though there is the simple round robin multiasking algorithm at work (provided tools like Executive are not being run) some tasks can hog the CPU. Now where do the task priorities come into play? Off the top of my head being that a task would be part of a node Exec would shuffle the task nodes in the task list so that the higher priority ones will get their time first before lower priority ones. As far as I understand it. :-)