This can be implemented externally, but here's an idea to make the estimates more precise: sleep for X seconds, measure the power usage, run the task, then subtract the background usage scaled to the duration of the task.
This can it easier to compare different implementations of a task.
This can be implemented externally, but here's an idea to make the estimates more precise: sleep for X seconds, measure the power usage, run the task, then subtract the background usage scaled to the duration of the task.
This can it easier to compare different implementations of a task.