I've just discovered Grid computing so please beare with me for wanting to learn a little more.

What i'm wondering if is the servers have a way to assign different tasks to systems?
Like would the scheduler have the ability to send "big" tasks to strong computers and "smaller" tasks to weaker computers? Or does it all just go after what the next in line is?

I do get that a strong computer would complete many small tasks faster than one big task. And that weaker computers would use long enough on small tasks and even longer on big tasks. But it seems somewhat more efficient to send small tasks to weaker systems and vice versa.

I do realize that this would depend heavily on what computers people do participate with. But is there any facotrs in play here at all?

The BOINC architecture allows the project to define "big" and "little" and assign appropriate work units. It is up to each project to decide whether their work would be better done using this feature.

I am not positive, but believe Rosetta has some tasks known to require high memory, and schedules them only to high memory systems.

---

Rosetta Moderator: Mod.Sense

Thank you for the quick answer. That does make alot of sense. I've heard before that biometrics scale better on high memory systems. Which is also why i asked