| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: Too many tasks on queue! |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 11
|
|
| Author |
|
|
wcgridmember
Advanced Cruncher Joined: Mar 30, 2005 Post Count: 112 Status: Offline Project Badges: 
|
I have 4 threads running at 60% on a laptop and I'm getting too many tasks to make calculations on. The symptoms: I have a total of 33 threads listed on the "Tasks" tab on Boinc Manager (4 of them running); that would not be a problem if the laptop would finish them all on time, but I've noticed that one of the tasks that is running should have finished by 16/10/2014 (4 days ago! - the task is Beta Test 7.22)! It's not any fun when we're trying to help mankind and because of an unnecessary tasks' mismanagement the work is the going down the drain...
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Abort the overdue beta task and many other members will stand in queue to compute it for you.
 If you think you have too many work units assigned, reduce the buffer either in the agent local prefs, or if you like via the website device profile. When you choose for latter, make sure to clear the local prefs, as these take precedence if they were ever created. |
||
|
|
wcgridmember
Advanced Cruncher Joined: Mar 30, 2005 Post Count: 112 Status: Offline Project Badges:
|
How do I do that?
|
||
|
|
asdavid
Veteran Cruncher FRANCE Joined: Nov 18, 2004 Post Count: 521 Status: Offline Project Badges: 
|
Web Site: select setting on the right up corner of the site, then from the left menu, device manager, then choose the profile associated with the device you want to customize. The parameter to change is Cache.
----------------------------------------Local prefs are found in tools menu on the boinc manager (computing preference)
Anne-Sophie
 |
||
|
|
wcgridmember
Advanced Cruncher Joined: Mar 30, 2005 Post Count: 112 Status: Offline Project Badges:
|
This is getting worse and worse. I now have 83 tasks at hand! I changed the cache earlier this day to 3 days, but after seeing this, I've changed it to 0,6 days. The beta task is running for almost 39 hours now. :-/
|
||
|
|
Falconet
Master Cruncher Portugal Joined: Mar 9, 2009 Post Count: 3295 Status: Offline Project Badges: 
|
Cancel the BETA.
----------------------------------------As for the cache, do you need a cache that high? 3 days is a lot. 0,6 is better. Set the cache to 0,2 and see if it improves. AMD Ryzen 5 1600AF 6C/12T 3.2 GHz - 85W AMD Ryzen 5 2500U 4C/8T 2.0 GHz - 28W AMD Ryzen 7 7730U 8C/16T 3.0 GHz |
||
|
|
Composer
Cruncher Joined: May 28, 2014 Post Count: 29 Status: Offline Project Badges: 
|
I don't mean to laugh at your difficulties, but for some reason I find the "this is getting worse and worse. I now have 83 tasks at hand!" very amusing. I also run asteroids@home, and at one point their servers wouldn't let me return tasks, but continued to send them to me. I ended up with 14 completed tasks sitting on my computer before it finally let me return the results in one massive deluge of scientific research... the whole experience for me seemed comically absurd...
|
||
|
|
wcgridmember
Advanced Cruncher Joined: Mar 30, 2005 Post Count: 112 Status: Offline Project Badges:
|
The Beta got to about 100 hours at a theoretical 100% done (which would be something like 99.999999999%) when my father killed it. The situation is now under control (0,15 of minimum buffer and 0,3 of maximum buffer).
----------------------------------------[Edit 1 times, last edit by wcgridmember at Nov 8, 2014 12:20:55 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Some tasks would somehow never get off the ground whilst usurping cpu time in full, never checkpointing and eventually dying by itself at maximum runtime exceeded or by the hand of fathers
 Not 100 percent sure, but think the last week released ugm version 7.23 coming out of beta fixed the issue.My android got 47+ fahv dump of repair tasks when plugging in on thursday night. Just letting them run sequentially in edf, earliest deadline first, and eventually will manually delete any that don't make it before deadline. Zero interested to let those near deadline compute in the know that the second time's up an extra copy is circulated. |
||
|
|
wcgridmember
Advanced Cruncher Joined: Mar 30, 2005 Post Count: 112 Status: Offline Project Badges:
|
Good to know that. Knowing that, it would be good to have BOINC kill tasks whose deadline has passed automatically.
|
||
|
|
|