| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: Too many workunits waiting to be processed |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 4
|
|
| Author |
|
|
BjornR
Cruncher Joined: Oct 9, 2009 Post Count: 1 Status: Offline Project Badges: 
|
Not too long ago i manually aborted tens of tasks because there was not enough time to process them in time, even if the system would've ran 24/7.
It hurts when having to abort workunits. I've set the work buffer to 0,01 days, yet i'm getting more than 100 workunits to be processed. 100 GPU workunits would be a mere 5 hours of work, 100 CPU workunits on the other hand would take 12,5 days to complete (as i'm typing this i realized that it processes 8 workunits simultaneously, but my point still stands. Boinc can't predict when i'm going to power the system off). Is there a way to define an absolute number of cached workunits instead of a variable amount based on time? Furthermore, is there a way to set the CPU and GPU limits separately? Most important system specs: CPU: Intel core i7 3820 @ 3.8 GHz GPU: 2 x MSI R7970 stock OC in crossfire setup I've read https://secure.worldcommunitygrid.org/forums/wcg/viewthread_thread,28636 |
|||
|
||||
|
PecosRiverM
Veteran Cruncher The Great State of Texas Joined: Apr 27, 2007 Post Count: 1054 Status: Offline Project Badges: 
|
Which ver of Boinc are you running?
----------------------------------------   |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
No there is no "how many units", except for CEP2, which has a special reason.
----------------------------------------Recommend to set Connect Every / Minimum buffer / Additional Buffer both to 0.00 days, then really there is no work fetch until a WU is complete. The version 7 client will anticipate the completion of a task and fetch a new task 1-2 minutes before the end. Running 8 GPU tasks simultaneous on a GPGPU is what you specified I'm sure with the 0.125 fraction. As for your other questions, no, no and no. None of those settings exist, nor is there separate buffer settings. The client already maintains separate scheduler for the 2 resources, prioritizing GPU over CPU when it comes to work assignment. P.S. The quota data is substantially outdated. The GPU allotment per day is in the thousands when deemed reliable. Edit: Typo per below post corrected. [Edit 1 times, last edit by Former Member at Mar 18, 2013 3:48:27 PM] |
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7691 Status: Offline Project Badges: 
|
prioritizing GPU over GPU when it comes to work assignment. I know there is a typo here, but which one is it ??????????????? Cheers Edit:corrected in above post.
Sgt. Joe
----------------------------------------*Minnesota Crunchers* [Edit 1 times, last edit by Sgt.Joe at Mar 19, 2013 2:26:08 AM] |
||
|
|
|