| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: Drug Search for Leishmaniasis Forum
Thread: Looong running WUs |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 117
|
|
| Author |
|
|
coolstream
Senior Cruncher SCOTLAND Joined: Nov 8, 2005 Post Count: 475 Status: Offline Project Badges: 
|
@etienne, here is a screencap for you. I hope it answers your question
----------------------------------------http://imageshack.us/photo/my-images/819/boinctaskscheckpoints.jpg/  Crunching in memory of my Mum PEGGY, cousin ROPPA and Aunt AUDREY. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
@etienne. The checkpointing doesn't seem to be every few seconds, but its also not as spread out as the sometimes several hours in between checkpoints in projects like CEP-2.
With the WU's I'm currently crunching, the longest time its been since a checkpoint is 15 minutes. |
||
|
|
armstrdj
Former World Community Grid Tech Joined: Oct 21, 2004 Post Count: 695 Status: Offline Project Badges: 
|
We are looking into the large running workunits for target 7. There is not an issue with results coming back only the sizing of the workunits. If you don't have a problem with long running workunits please continue to let them run the results are good.
Thanks, armstrdj |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi,
----------------------------------------I've done a short write up in the Checkpoint Saving FAQ in the Start Here forum. They are written at end every task inside the work unit. They can indeed take from very short to long, and the slower the machine, the longer it can take. My slow duo laptop is currently writing log entries every 34 minutes or so for a target 4. My [Linux] quad, stock speed, is doing a target 7 a bit quicker, every 7-8 minutes with a minimum setting of 5 minutes: 25509 WCG 06-09-2011 14:18 [checkpoint] result DSFL_00000007_0000004_0535_0 checkpointed 25511 WCG 06-09-2011 14:25 [checkpoint] result DSFL_00000007_0000004_0535_0 checkpointed 25513 WCG 06-09-2011 14:32 [checkpoint] result DSFL_00000007_0000004_0535_0 checkpointed 25515 WCG 06-09-2011 14:39 [checkpoint] result DSFL_00000007_0000004_0535_0 checkpointed 25518 WCG 06-09-2011 14:47 [checkpoint] result DSFL_00000007_0000004_0535_0 checkpointed 25520 WCG 06-09-2011 14:55 [checkpoint] result DSFL_00000007_0000004_0535_0 checkpointed 25523 WCG 06-09-2011 15:03 [checkpoint] result DSFL_00000007_0000004_0535_0 checkpointed 25526 WCG 06-09-2011 15:11 [checkpoint] result DSFL_00000007_0000004_0535_0 checkpointed 25530 WCG 06-09-2011 15:26 [checkpoint] result DSFL_00000007_0000004_0535_0 checkpointed Still it's projected to finish in 8.5 hours The client by default does not allow to write a checkpoint more often than once every 60 seconds, minimum, but it can't be forced. The application asks at each checkpoint if it is allowed to write one to disk. An aside, pity WCG chose not to use the same term as in e.g CEP2, "job", or "seed" or something, as tasks inside a task is confusing, but that could just be me. --//-- P.S. This morning's average validations had 8.09 hours, 33% over the objective of 6 hours mean. Suspect there's allot more variability in these target ligands than anticipated, which makes the cutting to size difficult. Hear that uplinger is looking at this issue. [Edit 1 times, last edit by Former Member at Sep 6, 2011 1:37:26 PM] |
||
|
|
coolstream
Senior Cruncher SCOTLAND Joined: Nov 8, 2005 Post Count: 475 Status: Offline Project Badges:
|
An aside, pity WCG chose not to use the same term as in e.g CEP2, "job", or "seed" or something, as tasks inside a task is confusing, but that could just be me. I agree that a common term would be easier to refer to. That's why I call them WUs (workunits) As an addendum to my previous link about checkpoints, I suspended WUs on an 8-core running Win7 after each WU reached checkpoint (wanted to make H/W fix). It took 7 minutes in all to suspend 8 units. Maybe I was lucky, I suspected it might take longer than that. EDIT: I forgot to add that these are all DSFL_00000007 WUs Crunching in memory of my Mum PEGGY, cousin ROPPA and Aunt AUDREY. [Edit 1 times, last edit by coolstream at Sep 6, 2011 2:42:13 PM] |
||
|
|
kffitzgerald
Senior Cruncher USA Joined: Jan 29, 2011 Post Count: 222 Status: Offline Project Badges: 
|
I'm running between 11 & 12 hrs for the 7 WU's I'm currently running on my Core I7 machine, will check what they started out laters when I get a few minutes
|
||
|
|
etienne06
Advanced Cruncher France Joined: Jun 11, 2009 Post Count: 56 Status: Offline Project Badges: 
|
Thank you Sekerob, coolstream and all of you who have tried to answer my question. I have a look to your chart & listing.
|
||
|
|
coolstream
Senior Cruncher SCOTLAND Joined: Nov 8, 2005 Post Count: 475 Status: Offline Project Badges:
|
Another good and informative post, Sekerob.
---------------------------------------- I still think it's a shame that BOINC Manager doesn't show checkpoints in the main screen (as BoincTasks does). Crunching in memory of my Mum PEGGY, cousin ROPPA and Aunt AUDREY. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Thanks for the complement.
The weaknesses of BOINC Manager has set the space for 1 man efforts such as by Fred to develop and maintain BOINCTasks+TThrottle. I hardly ever use the BM, certainly since they destroyed [IMHO] the tasks view and removed the message log tab. BT is enabled to easily set up to run on a Linux box too [in a WINE VM e.g.], but only use one instance, on the machine I work off [in Windows]. Hope he adds one day a screen to do do the cc_config.xml as once worked in the old BOINCView, if need be behind a secondary password to protect the unwitting change on shared devices. One of the main reasons to use BT is logging the completed tasks for both CPU and Elapsed time, so the efficiencies are on hand. DSFL on Linux does 99.2-99.6% when mixed with CEP2 i.e. it does not get cross-impacted in any significant way by the heavy duty Clean Energy jobs. --//-- |
||
|
|
uplinger
Former World Community Grid Tech Joined: May 23, 2005 Post Count: 3952 Status: Offline Project Badges: 
|
Sek,
I am looking into the estimated runtime per job and why the target 7 has caused such a large increase. Thanks, -Uplinger |
||
|
|
|