Hi,

I just installed BOINC on some Raspberry Pi boards and hope to use more in a class - instead of them all downloading and running different projects, which it seems to do as a default (?) can they all run the same project and get it done quicker?

Can everyone on a team also run the same project even though they have different email and passwords?

I am running SETI@home

Thank you!
Michael

Short answers, no and yes.

No, on the multiple devices running the same work. At Seti all tasks run on single hardware only, meaning one CPU = one task, one GPU = one task. There's also no multithreading applications, which would run one task with all the CPU cores the device has.

Everyone on a team can sign up with their own email address and password, furthermore, everyone runs their own work. You cannot run the same tasks as a team, unless one of your team is your wingperson (the other computer needed to return the work for it to validate).

---

Hi,

I'm new so I'm a little confused.

If the server puts together a task - let's call it "Task1" can I have my Raspberry Pi Zero, Raspberry Pi 3 and Raspberry Pi 2 all plugged in, all working on a part of that task?

Right now, I have a few Pi, all with my username logged in, but they all say 53+ hours to complete as they are different tasks...

If I had 5 Pi working on a 53 hour task together (Task1), I could do the task in just over 10 hours.

Thanks
Michael

BOINC doesn't work that way, as Ageless just said in the post before yours.
If you want tasks processed faster, then you need to get a single, faster, computer per task.

Most projects process single threaded tasks, serially, the present calculation step is dependent on the previous calculation step. Multithreaded are quasi parallel, were it not that there's still dependencies, so if one processor thread has done it's bit, the other might be waiting on the result before being able to continue. One member wrote about the multithreaded task cores were spending 90% waiting on other pieces to finish... highly inefficient. If then distributing parts of the calculation have to go across to other computer's CPU, that has even more latency with it, so does this in any way read as a good utilization of resources? Apart for the complexity added, not good of course, more playground than anything else.

---

Coelum Non Animum Mutant, Qui Trans Mare Currunt