| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: sudden change in WU completion times |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 9
|
|
| Author |
|
|
scrimenti
Cruncher Joined: May 15, 2008 Post Count: 7 Status: Offline Project Badges: 
|
Hi folks! This may not be a new thing to anyone else, but here goes...
----------------------------------------I'm so excited to be a part of this work that I am checking all my machines several times daily just to watch how my WUs cook along. (yeah, pathetic, huh?  ) Well, I've completed nearly 130 days worth of work, earned a bunch of badges, and now I recently noticed something new. My BOINC agent is requesting and getting tasks that run anywhere from 20 to 30 hours, instead of the 5 to 15 hours previously (but only on my best machine). Have I made it past probation and am now considered reliable enough for some heftier jobs, or is something amiss?Just curious. Scrimenti
âI am only one, but still I am one. I cannot do everything, but still I can do something; and because I cannot do everything, I will not refuse to do something that I can do." ~Helen Keller
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Different batches have different run times. We have had a few reports of long units, so it looks like we have some batches of big work units coming through.
So long as they complete and validate, there's nothing to worry about. |
||
|
|
scrimenti
Cruncher Joined: May 15, 2008 Post Count: 7 Status: Offline Project Badges:
|
We have had a few reports of long units... Whew, as long as it wasn't just me ! I thought something had happened to my CPU efficiency. I was a little sensitive to this, since one of my other machines locked up, and errored out in about 6 or 8 jobs just a day or so ago. (memory problems..no heartbeat from core) So long as they complete and validate, there's nothing to worry about. That's what I had hoped. Thanks! Scrimenti
âI am only one, but still I am one. I cannot do everything, but still I can do something; and because I cannot do everything, I will not refuse to do something that I can do." ~Helen Keller
|
||
|
|
JmBoullier
Former Community Advisor Normandy - France Joined: Jan 26, 2007 Post Count: 3716 Status: Offline Project Badges: 
|
Have I made it past probation and am now considered reliable enough for some heftier jobs? There is no such feature in WCG. Being considered reliable affects only the probablility for a given device to receive "emergency jobs" (also called "repair jobs"), that is WUs that WCG wishes to be returned fast and good because another WU of the same quorum has been returned in error or not returned at all before the deadline. Note that for this purpose a "reliable device" needs not only to be reliable but also to return its WUs quickly, i.e. within one or two days after distribution. One more comment: since WCG runs several projects with different durations per WU it is recommended that you mention the name of the related project when you report a problem or a question regarding the duration of a WU. Cheers. Jean. |
||
|
|
eviltoad
Senior Cruncher Australia Joined: Nov 5, 2005 Post Count: 190 Status: Offline Project Badges: 
|
I must admit I boggled a bit this morning when all my HPF jobs showed an expected duration of 37 hours, and one FAAH 46. Made me wonder if I should reboot my computer or something.
----------------------------------------Has anyone also experienced a sort of job pre-emption, where a pair of tasks would stop about halfway through and another pair start? I assume that once these re-prioritized tasks complete the paused ones resume, because they're usually gone by the next morning or when I get back from work or whatever.  |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello eviltoad,
Job pre-emption is a normal part of the scheduling process when BOINC thinks that there is danger of missing a deadline. Otherwise everything just runs in order. Lawrence |
||
|
|
JmBoullier
Former Community Advisor Normandy - France Joined: Jan 26, 2007 Post Count: 3716 Status: Offline Project Badges:
|
Yes, this is normal behavior. When Boinc detects that some task is at risk of missing its deadline it decides to run it High Priority and thus it has to stop one of the running tasks which becomes "Waiting to run". This task resumes as soon as one other task completes.
----------------------------------------In your case where FAAH and HPF2 WUs are involved it is possible that HPF2 WUs (which have a much longer deadline) will have to give way to the FAAH WUs which have become bigger these last days. Regarding the inflated duration estimates they should adjust - first, once an exaggerately inflated HPF2 WU has started its time to completion will decrease faster than usual - second, every time a WU completes faster than estimated the estimated duration of the pending ones will be lowered by up to 10 %. All in one, all this is normal and there is no need to reboot your computer (which should not improve anything anyway). Cheers. Jean. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
... My BOINC agent is requesting and getting tasks that run anywhere from 20 to 30 hours, instead of the 5 to 15 hours previously (but only on my best machine). ... Scrimenti Scrimenti, Can I ask about your system details that you are getting these wu's that take 20-30 hrs to complete? how fast is it? how much memory does it have? did you permit your system to put 100 % resource to the WU? |
||
|
|
Richard Mitnick
Veteran Cruncher USA Joined: Feb 28, 2007 Post Count: 583 Status: Offline Project Badges: 
|
I am so glad you are excited. I first started in when all i knew about was a Cancer project at Oxford University. One project, one site, but I was like a drunk person. To finally find something really useful to do with my computers besides work, music downloads, Wikipedia, etc.
----------------------------------------I came to WCG when I learned about it. Now I crunch on four machines, two projects at WCG and about 8 or so at BOINC. I am most excited by the Large Hadron Collider project at BOINC, well not at BOINC, nothing is "at" WCG or BOINC. LHC@home is the public project for C.E.R.N., US$8 billion, to find the Higgs Boson. So, glad you are here, keep up the fights, mankind will get somewhere. >>RSM |
||
|
|
|