| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: Human Proteome Folding - Phase 2
Thread: Invalid result after 100% completion |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 37
|
|
| Author |
|
|
knreed
Former World Community Grid Tech Joined: Nov 8, 2004 Post Count: 4504 Status: Offline Project Badges: 
|
Proposed a test to say e.g. if > 99.5% done, let it complete and pack up for transmission, but doubt it was heard by the developers. They have very selective hearing actually, because it was multiple times reported on their forums and possible a Trac ticket has been existing for longer. This has been discussed somewhat (if I remember correctly). However, apparently some projects do not correctly report their % complete and may sit at 99% done for many hours. I think the decision was that the BOINC client should behave 'fairly' regardless of an incorrect implementation by any project. [Edit 1 times, last edit by knreed at Jan 26, 2009 3:51:32 PM] |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Kevin, you realise this is a WCG project allowing BOINC to step away at 100% checkpoint. I've seen it too, but with 240 minutes project switch setting, there's a much greater chance WCG's time is not up right at 100%.
----------------------------------------
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
knreed
Former World Community Grid Tech Joined: Nov 8, 2004 Post Count: 4504 Status: Offline Project Badges:
|
Sekerob,
Sorry - I responded to a very narrow part of the discussion and didn't quote so my comments weren't clear. I've quoted what I responded to in my post. |
||
|
|
knreed
Former World Community Grid Tech Joined: Nov 8, 2004 Post Count: 4504 Status: Offline Project Badges:
|
dividedbymyself,
I only see one result returned by the K3 from you (it has returned more, but only one is left in the system). The file returned didn't contain any structures which is why the result was marked as an error. I would be curious to know if you got a different outcome for your results if you only ran WCG work (i.e. no project switching). I did look and there are a number of the same processor models and operating system/fixpack computers that are returning valid results for HPF2. However, most of them did look like they were dedicated only to World Community Grid so the potential for project switching to be an issue is there. If you are willing to just run straight WCG for a couple of days to help us understand if that is the source of the problem, we would appreciate it. thanks, Kevin |
||
|
|
dividedbymyself
Cruncher Joined: Aug 10, 2008 Post Count: 43 Status: Offline Project Badges: 
|
If you are willing to just run straight WCG for a couple of days to help us understand if that is the source of the problem, we would appreciate it. thanks, Kevin Hi again, I'm sorry for my absence over the past days. But I'm back again... Kevin, thanks for your investigation. Normally I run my computers for about 16 hours a day so my normal Boinc routine wouldn't "allow" to finish a WU of 25 hrs. in one run. But for the sake of the investigation I'll leave my computer running for a couple of days and suspend all other projects then HPF2. I think it will take another day from now before a new HPF2 Wu is downloaded. I'll suspend it to run that computer dry from my other projects, then start the HPF2 Wu to finish it in one run to see what will happen. And how many Wu's you'd like me to run in a row? Do you think two in a row would be enough? I'll report back in this thread as soon as the first is finished so you can check the result. Thanks for your attention, Bart |
||
|
|
knreed
Former World Community Grid Tech Joined: Nov 8, 2004 Post Count: 4504 Status: Offline Project Badges:
|
Go ahead and run at least one and we will see what happens.
|
||
|
|
dividedbymyself
Cruncher Joined: Aug 10, 2008 Post Count: 43 Status: Offline Project Badges:
|
Hi Kevin,
I've finished two HPF2 Wu's successfully according to the result status, both run without any interruption. At this moment I'm running a third that will be returned tomorrow (Monday), but during this one I've deliberately rebooted twice at random moments to check if this influences the result. After this one's finished I'll run another two while sharing time with one other project, changing project every two hours. The first I'll run without rebooting (unless I have to), during the second one I'll reboot a couple of times again when a HPF2 Wu is running. I guess these two will be finished around Thursday. After that I'm planning to continue the normal crunching routine of that computer again, but then one Wu will take about 2 weeks to finish again due to the resource sharing, unless you think it's useful to do some more testing. I'll keep you updated of any invalid result. Please let me know what you found out. Bart |
||
|
|
dividedbymyself
Cruncher Joined: Aug 10, 2008 Post Count: 43 Status: Offline Project Badges:
|
Kevin,
just to let you know that I've returned several Wu's without error now. The last few have been running in time-share with another boinc project (non-WCG). I'm returning to my normal crunching sequence again. If things go wrong again I'll report them asap so you can check it out. Thanks, Bart |
||
|
|
dividedbymyself
Cruncher Joined: Aug 10, 2008 Post Count: 43 Status: Offline Project Badges:
|
Kevin,
My last Wu went wrong again. It ended in a double "Boinc finished" call again after not finishing when the progress bar was at 100% and switching over to another Wu. It's strange that when I only crunched 2 or 3 projects (HPF2 being one of them of course) this didn't happen. I hope you can isolate and resolve the problem. If this keeps happening I maybe better concentrate on other projects that do not finish in an error. This is no threat, just a consideration from an efficiency point of view as I understand you have your priorities as well. Any advice? Bart |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Plz, can you confirm that the "Leave Application in Memory" option was on or off during the project swapping.
----------------------------------------To emphasize, if you only have WCG jobs in the queue you should not see swapping. Ah, yes, as per my previous recommendation, set the project switch time to 240 minutes to increase chance of jobs being finished directly when it's the HPF2's turn.
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
|