Credit Pending
12 Jan 2006 13:16:55 UTC
Topic 190591
(moderation:
No idea WHY some WUs credit despite being uploaded 10 days ago are still pending?!
BONIC seems to have created havoc with SETI as well. OH well we just keep crunching and the powers to be decide how the work is to be used or credited.
"Life is not measured by the number of
breaths we take, but by the moments that
take our breath away."
Regards; :-) Jim
Credit Pending
)
Jim,
We both come from the old school, so I'm sure you know all about reading the manual. There are two great documents easily available from the Einstein homepage. The BoincWiki is the most comprehensive, a labor of love by a very dedicated volunteer. The frequently asked questions has the answer to you exact question. ;-)
Regards,
Michael
microcraft
"The arc of history is long, but it bends toward justice" - MLK
It really irritates me when
)
It really irritates me when people state that BOINC has screwed the pooch, now just how did BOINC effect pending credit???? What problems at SETI????
Credits are pending because you need a certain number of successful WUs to reach a quorum, then they are validated, see BOINC has no control over this. process.....I am finished.......
BOINC Wiki
not sure if it helps, but
)
not sure if it helps, but look at your pending results and click on the work id. there you can see if someone is taking forever to report back. boinc (?) waits till there are 3 results for the same unit. 2 users are holding up 6 results for me.... usually the people that overque themselves.
The fact there's *any*
)
The fact there's *any* pending that's been outstanding for a while, considering the scope of the project, suggests a pretty high mortality ... or low probability of completion on any given host in a short period of time.
Albert doesn't have any
)
Albert doesn't have any redundancy in work allocation, so only a single person is needed to delay a WU for an extended ammount of time. I keep a 4day queue because my ISP doesn't appear to have an evening/weekend sysadmin, this means that I'm generally the last person to submit, I've had about a dozen WUs stuck for 2 weeks because one of the people doing them dropped after only doing a few days of work.
I might drop back to 3 days since we're past the extended holiday weekends, the 4th day covers me against a problem requiring hardware replacement instead of a simple reboot so I'm hesitating.
One thought that's occured to me is that until someone's done 2 weeks of reliable work only count them as one half in the scheduler. This would largely remove delayed credit when someone quits after a day with 2 weeks of work pending.
sorta; its the people who
)
sorta; its the people who overque. i looked at a result that was pending and the holdup was a user that i finally found... it was on the THIRD page... meaning they had downloaded over 60 units to compute... even running 16 reports a day, thats a lotta days of work. the procedure i usually use is to click on your results. look for pending ones, specially old pending ones and click on the work unit id and it will show you the 3 people who are computing on that particular unit. one person probably hasnt completed their work; copy the and copy the result id from the left column (from their line), click on computer in the right column which will give you a computer summary; click on results (near the botom) and search for the result id. phew- sounds long but its just follow the link tag. one i just checked has 9 units in front of it. at 1.5 hrs a unit its only a 3/4 day wait if hes running 24/7 but if he decides not to report for a few days or only computes say 6 hours a day, i am in for a 2.25 day wait for credit. you have to go to your account and edit preferences to download a lot; they discourage getting more than 10 at a time. i adjusted my prefs to i think connect to network every .85 days and it only gets 7 or 8 then i change boinc prefs to no new work, crunch and report till all of its done then do it all over again. i will agree with the within a short period of time part
>The fact there's *any* pending that's been outstanding for a while, considering >the scope of the project, suggests a pretty high mortality ... or low probability >of completion on any given host in a short period of time.
Not just overque. My pending
)
Not just overque. My pending suggests a large number than download a few, maybe complete 1 or 2 (if any), then disappear. Maybe aliens got 'em.
RE: Not just overque. My
)
glad I ain't the only one suffering with those people.
RE: RE: Not just overque.
)
Can I just repeat that, i've got 12 D/L'd on 2nd and matched with 2 or 3 people that I don't believe are going to do the units.
Everyone, This is really
)
Everyone,
This is really unremarkable, "luck of the draw" and nothing else. So, they timeout, get re-assigned, crunched and credited, no problem. It's a simple fact that there's a high drop-out rate for noobs - probably 25% don't make it through the first 2 weeks. It's that way across the projects, has been that way throughout, and the trend's not likely to change. I saw the same thing back in July. My "pending" list grew to 35 or more, Einstein WUs, not these short little alberts, so it was about 200 hrs worth of my machine time, over a week at 24/7, just lying there. They were re-assigned, re-sent, and I wound up with a huge bump in RAC in late-July early-August, 600- and 700-credit days. Too bad for me??? No, I got the credit in good time and it's still part of my score. Only possible victim was the science that was delayed.
Michael
microcraft
"The arc of history is long, but it bends toward justice" - MLK