Diapositive 21 sur 26
We can imagine a lot of cost notions. If we just consider the convergence time it means we think it is exactly equivalent having 10 or thousand people working on a project as soon as the aim is reached on the same time.
It might be right for very particular tasks, where time is critical, but usually, it is not the case. A more realistic cost is the second one. Nevertheless, it means something only if we use the same algorithm. We can not compare two algorithms by using t
That is why I think processor time is finallyquite good, unless you have a better idea. Of course, now, it depends on the machine and on the language. That is why, finally, I evaluate for each run the number of "basic" operations, such as addition, multip