I'm having an issue with N-body .40 (mt) taking over my system, its downloading a WU, assigning it a nonsense "to completion" of about 58 hours, punching into high priority, finishing in about 40 mins then DLing another WU to do it all again.
The only non-standard apps I'm using in BOINC is SETI optimized apps.

Any thoughts?

Yes, they are having some problems getting the new N-Body re-integrated in the grand scheme of things since the big project upgrade last month.

In your case it sounds like you are having Duration Correction Factor issues. So if the host is completing the N-Body tasks successfully, my suggestion is to just leave it alone and let BOINC make the DCF correction on its own.

Once it has gotten that straightened out, it will pay back all the extra time it runs up now in High Priority to your other CPU projects eventually.

I've been having the same problem for the past few weeks, and it just got insanely worse. My latest MilkyWay@home tasks have started out claiming they needed over 3000 hours of my computer's time for the next 6-8 days, which is manifestly not possible, then after sucking all my available CPU (suspending all my other BOINC tasks), they finish in only 3+ hours. Of course, then the next one downloads and takes over again.

I haven't seen any factor corrections made (unless they went the wrong way!), so unless someone has a better way to fix this, I think I'll just refuse new tasks until the N-body tweaks are done.

on one of my host the all N-body .40 mt WUs is ending with this error :

There are no child processes to wait for. (0x80) - exit code 128 (0x80)

what happen ?

---

I quite like the n-body WUs. They run quite fast on my machine (under an hour I believe but I usually don't see them actually complete).

Is there a way to only run n-body on my cpu and de_seperation on my gpu? Einstein@home has a nice option in the preferences which you can tick so you only get cpu work that isn't available for your gpu too.