Seti@Home optimized science apps and information

Optimized Seti@Home apps => Windows => Topic started by: funonline on 25 Sep 2006, 10:53:42 am

Title: Computing forever for some WU´s
Post by: funonline on 25 Sep 2006, 10:53:42 am

Hi @ all,

I´ve see it now two times ...  Seti @ Home deliver workunit´s who are computing "forever".
The first time I saw this, I cancel the WU ... so far so good ...
But now I´ve got the second WU who had the same behavior  >:(
The WU "stands" ... no futher progress will be made, no matter how much time you will give it.
Only the original client could compute it, the first faulty WU finished only by original clients.
So I think, this must be a little bug in the optimized client ?!
BTW : You see no CPU usage when the WU stops ...

I´m using the optimized app version 5.15
WU´s:
From the first one, I don´t have any links remaining  ???
But the second one is :
http://setiathome.berkeley.edu/workunit.php?wuid=92236319

If you want I could supply you with the WU and maybe traces of it.
Thanks for any help ! :)
CYA
Funonline

EDIT : WU attached ... it stops about 66 % / 67 %

[attachment deleted by admin]

Title: Re: Computing forever for some WU´s
Post by: Simon on 25 Sep 2006, 11:15:36 am

Hiya,

sure, you can attach a zip file of the WU in question - I'll try and run it on one of my hosts and see what happens. Personally, I've never had a "stuck" WU that I've noticed, but then I don't have time to watch my crunchers too closely, so I can't say for certain.

Regards,
Simon.

Title: Re: Computing forever for some WU´s
Post by: funonline on 25 Sep 2006, 11:49:56 am

Hmmm, now it´s becoming strange ...
I´ve shutdown my boinc-client to zip the file.
After that I start it again.
The WU was reseted to 51 % and computes again ...
Now I´m over 66 / 67 % and still computing ... strange ...
Maybe I could finish it now ?

CYA
Funonline

EDIT:
I finished the WU after the restart ...

Title: Re: Computing forever for some WU´s
Post by: hiamps on 30 Sep 2006, 02:32:19 am

I had that problem a couple times before Simons apps were released. Shutting down Boinc and restarting worked both times. There was a thread about it on the seti board.

Title: Re: Computing forever for some WU´s
Post by: archae86 on 02 Oct 2006, 08:50:04 pm

I've had several WU's "stall out" in the last few days, running KWSN.  I'm running a Gallatin hyperthreaded, with SETI set to a fairly low percentage with Einstein getting the rest.

The behavior of "bad" WU's is that the client will give a half hour time slice to SETI, but once the bad behavior sets in that SETI science ap will rack up zero additional CPU time.  System Idle then gets 50% of CPU time. The Einstein side will compute on normally, undisturbed, and the CPU temperature drops about 2 degrees C.  When the half hour time-slice for SETI ends, the second Einstein job restarts normally, temperature rises back, idle goes to zero, and all is well.

So far I've just aborted the WU's when I've noticed.  If I see it again, I'll try stopping and restarting BOINC as suggested in this thread.