Completed BRP4 (Arecibo) Tasks 48+ hours ago - Still "Unsent" to Wingman
7 Aug 2013 19:02:03 UTC
Topic 197122
(moderation:
I'm not sure if this is the correct forum to report this, but I'm showing three of my BRP4X64 (Arecibo) tasks as "Pending", two of which haven't even been sent out to a wingman over two days after I returned my results. WTF?
Completed BRP4 (Arecibo) Tasks 48+ hours ago - Still "Unsent" t
)
I wouldn't worry about that since it isn't very much credit and I see those Androids have been getting those tasks.
But since you have no tasks even running right now maybe you should use your GPU card and run some Binary Radio Pulsar Search (Perseus Arm Survey) tasks.
You will have better luck doing that.
RE: RE: I'm not sure if
)
Let me try this again ...
The two tasks I cited haven't been sent to ANYONE, Android or otherwise ...
The reason that I have no tasks running right now is quite simple. I wanted to try a couple of small tasks at first, to see how much lag there would be, before getting credit applied. Wasting three or four hours of computer time as a test made more sense than winding up wasting three or four weeks.
I know it's a small sample, and things happen -- but it's certainly not a very auspicious beginning, and an even less auspicious response to my problem report.
Must admit I've never seen
)
Must admit I've never seen that before, and will be interested to know why. Maybe there's a lack of suitable wingmen?
Currently the BRP4 search is
)
Currently the BRP4 search is running on "low priority mode". As in, the server will more likely send tasks for the new S6casa search or the FGRP2 to volunteers.
So no bug there, your work units will be assigned to a wingman. Just taking more time than the usual.
I thought this Not sent as
)
I thought this Not sent as unusual...
For the last few weeks i have - for hot weather reasons mainly - stopped requesting cpu tasks in the E@H preferences page. (as an aside some gpu tasks run 10% quicker with no cpu load)
The server page says there is plenty to send.
I´m running Linux, (and it has cooled down!), i could(?) be a valid wingman, so i have just tried updating the preferences and project to pick up some BRP4.
I should pick up some BRP4 tasks.... if not the ones in question.
Wed 07 Aug 2013 22:33:15 BST Einstein@Home Sending scheduler request: To fetch work.
Wed 07 Aug 2013 22:33:15 BST Einstein@Home Requesting new tasks for CPU
Wed 07 Aug 2013 22:33:45 BST Einstein@Home Scheduler request completed: got 0 new tasks
Wed 07 Aug 2013 22:33:45 BST Einstein@Home Message from server: No work sent
Wed 07 Aug 2013 22:33:45 BST Einstein@Home Message from server: see scheduler log messages on http://einstein.phys.uwm.edu//host_sched_logs/4918/4918234
Wed 07 Aug 2013 22:33:45 BST Einstein@Home Message from server: No work available for the applications you have selected. Please check your preferences on the web site.
Seems not.
Or am i missing something?
Perhaps I am - the scheduler
)
Perhaps I am - the scheduler log has
2013-08-07 21:33:45.6545 [PID=10964] [send] [HOST#4918234] [WU#171164218 p2030.20121022.G194.79-01.99.C.b4s0g0.00000_1473] WU is infeasible: BRP4 work currently limited to mobile devices
so unless i put some wheels on my device it can not be a wingman in BRP4!
Surprise surprise, BRP4 is
)
Surprise surprise, BRP4 is now alive - but not as we know it Jim. Well they look odd to me.
Some strange tasks delivered...
http://einsteinathome.org/workunit/169342346
has a matching Unsent WU
This guy
http://einsteinathome.org/workunit/169310898
has 4 in progress WUs and one unsent WU.
and
http://einsteinathome.org/workunit/168842589
has 2 Unsent. etc.
Or are these tasks exceptional ones for some reason?
RE: ... three of my BRP4X64
)
Tasks remain 'pending' until the second copy is sent out and returned. There is nothing unusual in what you are seeing.
This doesn't happen very often but it can (sometimes) and it's not a problem when it does. There can be a variety of reasons for a delay in sending out the second copy of a task. I would simply observe them for a few more days and raise a further query if the second copy hasn't been sent out after (say) another week.
Cheers,
Gary.
RE: Some strange tasks
)
If you look closely at each of the three workunits you listed, you will find that the 'initial replication' has been increased from 2 to a higher figure.
This would have been a deliberate decision made by the Devs to get a small number of critical workunits completed as quickly as possible. Unfortunately, quite a few people download tasks and never return them. In runs that take many months, this isn't too much of an issue as there is plenty of time to resend these tasks - more than once if necessary - after the normal 2 week deadline has passed.
In this particular case, we are talking about a special batch of the very latest data from Arecibo, that the project was able to get access to before 'other channels'. It was sent out with great urgency - I think it lasted about a week only. Consequently, it's not surprising that there is some urgency in 'cleaning up the dregs'. You should note that in each of your examples there was a 'deadline miss'. Perhaps it's the 'deadline miss' that triggers the increase in initial replication, which then results in the extra tasks being put into circulation.
Cheers,
Gary.
RE: RE: ... three of my
)
Thank you for a reasonable response, Gary. I do understand what "pending" means in the BOINC universe, though :)
I guess I've been spoiled by other projects like SETI or WCG, where the time between the two (or more) instances sent is more usually measured in seconds or minutes, not days.