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: 58
|
![]() |
Author |
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7697 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
BOINC Manager at least should be able to adjust its estimates. When a task shows 50% completed with only 8 minutes elapsed, and the estimated time remaining is 2 hours, it shouldn't take a lot of AI to make a correction. I wholeheartedly agree on that point. Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
asdavid
Veteran Cruncher FRANCE Joined: Nov 18, 2004 Post Count: 521 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
BOINC Manager at least should be able to adjust its estimates. When a task shows 50% completed with only 8 minutes elapsed, and the estimated time remaining is 2 hours, it shouldn't take a lot of AI to make a correction. I wholeheartedly agree on that point. Cheers This is something i noticed when installing the new version of the manager (7.2.47). It took much more time to adjust remaining times than the old 6.x.x one.
Anne-Sophie
![]() |
||
|
Crystal Pellet
Veteran Cruncher Joined: May 21, 2008 Post Count: 1323 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
BOINC Manager at least should be able to adjust its estimates. When a task shows 50% completed with only 8 minutes elapsed, and the estimated time remaining is 2 hours, it shouldn't take a lot of AI to make a correction. I wholeheartedly agree on that point. Cheers This is something i noticed when installing the new version of the manager (7.2.47). It took much more time to adjust remaining times than the old 6.x.x one. World Community Grid decided not to use BOINC's duration_correction_factor and therefore that factor is always 1.000000. So BOINC fully depends on the estimated time for the tasks that it get from the WCG-project. CP |
||
|
rbotterb
Senior Cruncher United States Joined: Jul 21, 2005 Post Count: 401 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
If you find your running less than a full number of WUs for your system, you can always in a short term change the perferences to a larger number and force a task update. You'll get can usually get a bunch of extra WUs to crunch while afterwards you put your perference level back to where it was before. Now this if obviously rather manual, but it can get more WUs to your system if you want them (and assuming there are WUs out there to download in the first place).
|
||
|
l_mckeon
Senior Cruncher Joined: Oct 20, 2007 Post Count: 439 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
If you find your running less than a full number of WUs for your system, you can always in a short term change the preferences to a larger number and force a task update. You'll get can usually get a bunch of extra WUs to crunch while afterwards you put your preference level back to where it was before. Are you sure? While I haven't tried changing my preferences, with a fairly empty queue and forcing an update, I think I still only get three tasks at a time. |
||
|
sean0118
Cruncher Australia Joined: Feb 7, 2010 Post Count: 33 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Do you mean under special circumstances? I just increased my cache a bit and got 16 new tasks.
|
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
with a fairly empty queue and forcing an update, I think I still only get three tasks at a time. If you double your minimum cache buffer setting, you should get more than just the three. |
||
|
PMH_UK
Veteran Cruncher UK Joined: Apr 26, 2007 Post Count: 774 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Are you sure? While I haven't tried changing my preferences, with a fairly empty queue and forcing an update, I think I still only get three tasks at a time. You may still only get 3 per call but can update again for more, up to buffer prefs. Tasks per call seems to differ per BOINC level and CPU speed. Paul.
Paul.
|
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Iirc, the newer agents will fetch a task or two even if work is not needed, does say so if that's the case in the event log! This is simply to not let a server rpc call go to waste. How often you can do this hitting of 'update', until the agent tells you to go climb a tree i do not know. The cpdn project have for those heavy keyboard pounders a 3636 second enforced backoff. The wcg 121 seconds is by comparison benign, but still a whole lot more than the 7 seconds that simap tells you to exercise patience.
For those not afraid to read, i mean read a few more log lines, you can add the sched_op_debug tag to the cc_config.xml. Then when hitting update you get actual feedback on how many seconds of work are requested and how many seconds are given. Standard seconds as it is as noted based on a frozen dcf of 1 for version 7 agents. 2751017 Default 106764 World Community Grid 6/28/2014 4:08:39 PM update requested by user 106765 World Community Grid 6/28/2014 4:08:43 PM sched RPC pending: Requested by user 106766 World Community Grid 6/28/2014 4:08:43 PM [sched_op] Starting scheduler request 106767 World Community Grid 6/28/2014 4:08:43 PM Sending scheduler request: Requested by user. 106768 World Community Grid 6/28/2014 4:08:43 PM Not requesting tasks: don't need (zero resource share) 106769 World Community Grid 6/28/2014 4:08:43 PM [sched_op] CPU work request: 0.00 seconds; 0.00 devices 106770 World Community Grid 6/28/2014 4:08:46 PM Scheduler request completed 106771 World Community Grid 6/28/2014 4:08:46 PM [sched_op] Server version 701 106772 World Community Grid 6/28/2014 4:08:46 PM Project requested delay of 121 seconds 106773 World Community Grid 6/28/2014 4:08:46 PM [sched_op] Deferring communication for 00:02:01 106774 World Community Grid 6/28/2014 4:08:46 PM [sched_op] Reason: requested by project |
||
|
l_mckeon
Senior Cruncher Joined: Oct 20, 2007 Post Count: 439 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
The point was if you've got an empty queue and they're only sending you three at a time, it can take half an hour or more to restock with short WUs. Bad design, and it didn't used to do that.
|
||
|
|
![]() |