| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: The Clean Energy Project - Phase 2 Forum
Thread: Question about valid tasks |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 16
|
|
| Author |
|
|
svincent
Advanced Cruncher Joined: Jan 3, 2009 Post Count: 53 Status: Offline Project Badges: 
|
I agree with littlepeaks: it does seem that bad work units have negative consequences beyond the time wasted on that task. A host that gets such a task will presumably no longer be deemed reliable and future work units, at least for a time, will require a wingman, resulting in inefficient use of resources. This is consistent anyway with what I'm seeing in the Results Status list.
|
||
|
|
Yarensc
Advanced Cruncher USA Joined: Sep 24, 2011 Post Count: 134 Status: Offline Project Badges: 
|
That is how it works for CEP tasks, according the Sekrob in this post, it takes 20 valid results to not require a wingman again.
http://www.worldcommunitygrid.org/forums/wcg/...ead,38082_offset,0#493236 But on the positive side, that work unit not even being able to reach the first checkpoint helps the scientists and techs better understand what limits to put in place to avoid that happening in the future. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
-Never- seen a task not make it to the first checkpoint and got 4600 validated CEP2 tasks to my name, but for 1 day short of 4 years. I do see tasks at times not managing to complete within the 18 hours, but then credit is proportionally given if there is a wingman, and 'claimed' is given if it ran alone.
If not making first checkpoint within 18 hours on a regular basis, then maybe the device has to be opted out. Technically, the heavier tasks are already filtered out, but the method is not ironclad. A weighty one does get through now and then. If it hits on a powerhouse-node, no issue. If hitting on a Centrino, unlucky. Drawing a horizontal line on this chart http://bit.ly/WCGCEP2 , current mean is 7.5 hours, long term mean is about 8.5 hours, and peak mean was just over 9, ergo any descent host [that was opted in] is expected to make it to first checkpoint. A bad unit in amongst... yes can happen. It's like petri dishes in a lab... some do not show the wanted result. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I'd like to know what last job consists of since some other jobs have been briefly explained already but not this one. Is the explanation at 31:30 in this video the answer to my question? |
||
|
|
Enchanted
Cruncher Joined: Aug 29, 2010 Post Count: 4 Status: Offline Project Badges: 
|
Is the explanation at 31:30 in this video the answer to my question? You're indeed curious. :) Some details on the current WU layout may be found here. I'm not sure about RC=0x1, but my guess is that the last job may sometimes be impractical to compute (e.g. there's only one "good" functional). |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Thank you Enchanted for the link, it's an exausive post and it's not easy to find it in the whole forum.
----------------------------------------I'm used to projects whose teams put a certain effort in answering questions from crunchers instead of answer vaguely so my curiosity about how WUs work grew up. It's a way to keep people interested in the project which fuels their will to help in their own small way with their own little pc and few money and crowd computing is based on this. About return code 1 I've googled and found out that usually it appears when some errors in writing the application have been made (privileges, location of certain files), but reading this forum it seems it pops out when WUs which are too tough for humble PC are sent out to crunchers instead of remaining in the in-house cluster. But since I'm not fond of all this I don't exclude your explanation too and I thank you for it. EDIT 16/8 A new return code popped out, valid WU. Result Name: E232263_ 23_ S.310.C25H13N21.JICQIKCZJHKMFS-UHFFFAOYSA-N.4_ s1_ 14_ 0-- <core_client_version>7.4.42</core_client_version> <![CDATA[ <stderr_txt> INFO: No state to restore. Start from the beginning. [10:00:51] Number of jobs = 8 [10:00:51] Starting job 0,CPU time has been restored to 0.000000. [12:45:10] Finished Job #0 [12:45:10] Starting job 1,CPU time has been restored to 9767.343750. [13:16:10] Finished Job #1 [13:16:10] Starting job 2,CPU time has been restored to 11619.859375. [13:25:26] Finished Job #2 [13:25:26] Starting job 3,CPU time has been restored to 12169.453125. [13:36:24] Finished Job #3 [13:36:24] Starting job 4,CPU time has been restored to 12824.218750. [13:45:26] Finished Job #4 [13:45:26] Starting job 5,CPU time has been restored to 13362.531250. [13:52:50] Finished Job #5 [13:52:50] Starting job 6,CPU time has been restored to 13796.843750. Application exited with RC = 0x3 [15:21:24] Finished Job #6 [15:21:24] Starting job 7,CPU time has been restored to 18942.765625. [15:21:24] Skipping Job #7 15:21:32 (6696): called boinc_finish </stderr_txt> ]]> [Edit 1 times, last edit by Former Member at Aug 16, 2015 1:58:25 PM] |
||
|
|
|