| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
| World Community Grid Forums
|
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 30
|
|
| Author |
|
|
Bearcat
Master Cruncher USA Joined: Jan 6, 2007 Post Count: 2803 Status: Offline Project Badges:
|
Running Ubuntu lucid with all updates. Boinc is version 6.10.17. When bionic updates and gets wu's with an earlier return time than what's in my pending, bionic seems to go right down my list and not pay attention to the ones with sooner deadlines. Seems to fit them in periodically. Any way to force bionic to prioritize?
----------------------------------------
Crunching for humanity since 2007!
![]() |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
runs them in order downloaded, until they get in danger of being late, then they jump to the top.
let me guess, is HCC wu with the early date?? they just recently changed it from 10 to 7 day deadlines, having same problem. as long as cache is not to big, they will process in order and on time |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
What fredski said, processing is in FIFO order First In First Out unless deadline panic breaks out.
----------------------------------------For those on a high cache of > 6 days the HCC tasks will get their prioritisation, the client switching at times to EDF, Earliest Deadline First. Those running only HCC with a cache > 6 days could be seeing their client in permanent panic state, high priority processing, and if over 7 days is being cached, overdue work eventually being auto-cancelled. The techs wrote in that they were on storage cram. 3 Days production less of tasks listed on the Result Status page means about 600,000 less tasks on the live system. Maybe (shear speculation), after all the upgrade/conversion work is complete they'll extend the deadline again, but think not. The grid is growing rapidly heading for 1 million tasks completed per day or 11.57 per second.
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Bearcat
Master Cruncher USA Joined: Jan 6, 2007 Post Count: 2803 Status: Offline Project Badges:
|
Running clean water.
----------------------------------------
Crunching for humanity since 2007!
![]() |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Bearcat,
----------------------------------------Presume you understood the replies so far of FIFO operation? On your question: Any way to force Careful what you wish for, yes it can be done, but it comes at a prize!!!! The trick is to set the Switch Between Application Every xxx Minutes to for instance 6000 minutes... higher than the deadline. All these jobs will jump the queue. The prize is that the jobs with normal WCG deadline will be suspended. If LAIM is on they will sit in memory. If LAIM is off the suspended tasks unload and resume on first opportunity from last checkpoint. Works for regular repair jobs and short deadline Beta tasks, so much so that WCG will see the device as a high speed returner and send even more to the point that you can have an octo with 8 jobs processing in high priority whilst 8 others are sitting waiting. Done it on my quad and got frustrated, so set it back to only do this when there's Beta and DDDT2 Type A in the air... both might be soon. Your choice. Yes I know, it's "price" ;>)
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
jsobry@gmail.com
Cruncher Canada Joined: Jan 11, 2008 Post Count: 31 Status: Offline Project Badges:
|
I have a similar problem and find the updates here too cryptic. Can someone explain how we pick the next task and how we can fix this problem. Right now I intervene and suspend and then resume various tasks. The current procedure is to run the tasks that get late at high priority but that does not have any effect on an 8 way processing system running 8 tasks concurrently. Eventually the late tasks run out of time and all efforts come to naught .That is a situation we should avoid I think.
----------------------------------------[Edit 1 times, last edit by jsobry at Oct 19, 2010 5:59:18 PM] |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
"I have a similar problem and find the updates here too cryptic."
----------------------------------------Agree your line is cryptic and absent detail information a big guess to serve you with tailored advise. Reduce your cache to below the shortest deadline of any project your participate in and even if you don't want to do that, it's best to let BOINC manage the processing sequence and you stop knobbing the order. Takes several weeks before the client has it figured out, particularly if you also crunch outside of WCG.
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 1 times, last edit by Sekerob at Oct 19, 2010 6:09:09 PM] |
||
|
|
Ingleside
Veteran Cruncher Norway Joined: Nov 19, 2005 Post Count: 974 Status: Offline Project Badges:
|
I have a similar problem and find the updates here too cryptic. Can someone explain how we pick the next task and how we can fix this problem. Right now I intervene and suspend and then resume various tasks. The current procedure is to run the tasks that get late at high priority but that does not have any effect on an 8 way processing system running 8 tasks concurrently. Eventually the late tasks run out of time and all efforts come to naught .That is a situation we should avoid I think. You didn't say which BOINC-client-version you're using, but many of the v6.10.xx-clients (including v6.10.17 if not totally mis-remembers things) has various scheduling-bugs there the wrong task is selected to run as high priority... So, if you're using an older v6.10.xx-client, it's a good idea to upgrade to the current version, v6.10.58. ![]() "I make so many mistakes. But then just think of all the mistakes I don't make, although I might." |
||
|
|
Bearcat
Master Cruncher USA Joined: Jan 6, 2007 Post Count: 2803 Status: Offline Project Badges:
|
How solid is 6.10.58? Hate to fix something that's not broken. For now, I will reduce my cache until it's caught up.
----------------------------------------
Crunching for humanity since 2007!
![]() |
||
|
|
jsobry@gmail.com
Cruncher Canada Joined: Jan 11, 2008 Post Count: 31 Status: Offline Project Badges:
|
I downloaded the latest version 6.10.58 on all my PCs and Macs. Still have the same problem. I only run WCG.
----------------------------------------Question: how does the server decide which task to download? It looks like we get new tasks with an earlier deadline than the ones we are already running. Is that the case? [Edit 1 times, last edit by jsobry at Oct 20, 2010 6:18:24 PM] |
||
|
|
|