Overcommitted
30 Dec 2005 17:28:29 UTC
Topic 190473
(moderation:
Each time I allow Einstein and Seti to get new work, BOINC downloads 8 jobs each. With so many jobs, the program is overcommitted and just works on the one project instead of distributing the work equally between Seti, Einstein, LHC, and Climate. What changes can I make to prevent this?
Overcommitted
)
there's a few things we will need to know to give you a definitive answer. First of all, what is your "connect to.." setting? Perhaps you could lower it? When you say "Each time I allow E and S to get new work..." what method are you using to stop them from getting new work? How long does this over committed state last? Oh... what kind of machine are you using? (ie speed, memory, OS, etc)... Answers to these and other questions will help locate and correct the problem.
Jim
PS... Hello Jord. I just poked my head in to see what was going on. I thought I might be able to get something moving on this request. Laters...
Jim
Also need to know the version
)
Also need to know the version of BOINC Manager, and if the CPDN WU is a "slab" or a "sulphur", and what the deadline is on it, and current % done...
But main thing is probably just to lower your cache size, adjust your resource shares to what you want for SETI/Einstein/LHC/CPDN, take them _off_ "no new work" or "suspend", (well, actually, if LHC has no work, it's better to "suspend" it rather than have it constantly trying...) and let BOINC handle the scheduling. When you try to micromanage it, you _cause_ "overcommitted" more often than not.