| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: Human Proteome Folding - Phase 2
Thread: Long-running HPF2 task |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 39
|
|
| Author |
|
|
schepers
Advanced Cruncher Canada Joined: Oct 11, 2006 Post Count: 85 Status: Offline Project Badges: 
|
I was spot-checking one of my systems (Intel C2D 2.16Ghz, BOINC ID 94530), and noticed one of the two active work units was HPF2, but was already at 28 hours, with an estimated 12 hours to go. Since all the other HPF2's were averaging 3-4 hours, and I wasn't aware of any monster tasks like this for HPF2, this seemed like an anomalous task and canceled it.
Task: HPF2 v5.18 Task name: ZA255_00079_15 Downloaded: Feb 20 @9:13 AM (EST) Task started: Feb 22 @3:27 AM (EST) The system is now running two HPF2's in parallel and they appear OK. Any ideas as to what was happening? |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
No, but on occasion a particular one will not find the solution easily and will go on much longer. If CPU % is still normal, graphics moving and CPU time still accumulating, it's not broke yet.
----------------------------------------Suggest to check the Result Status detail quorum. If all are finished in a short period and points have been awarded, u got a crooked one which is best aborted. Oh, before i forget, running HPF2 and HCMD on UD agent side by side with BOINC or similar low priority process, will grind the UD to trickle progress. Deleting the Throttle tool registry entries and using the environment var. will let them move faster.
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 1 times, last edit by Sekerob at Feb 23, 2007 1:46:17 PM] |
||
|
|
schepers
Advanced Cruncher Canada Joined: Oct 11, 2006 Post Count: 85 Status: Offline Project Badges:
|
I checked the faulty task return results, and most of the others had reported in with run-times ranging from ~3-9 hours. I think this confirms that mine task was faulty. When I canceled it, the claim was already up to 366, with the actual granted claim for the returned tasks being 46.2. Mine would have completed (after wasting 2 days) and have been rejected as either "error" or an out-lier. I'm glad I caught it and canceled it.
|
||
|
|
cio_redulla
Advanced Cruncher Philippines Joined: Apr 24, 2006 Post Count: 130 Status: Offline Project Badges: 
|
I checked the faulty task return results, and most of the others had reported in with run-times ranging from ~3-9 hours. I think this confirms that mine task was faulty. When I canceled it, the claim was already up to 366, with the actual granted claim for the returned tasks being 46.2. Mine would have completed (after wasting 2 days) and have been rejected as either "error" or an out-lier. I'm glad I caught it and canceled it. I also experienced having these long work units. I restarted BOINC and it continued crunching, it didn't take that long however, it only took a little more than two hours (from the projected 3-9 hours). cio_redulla  |
||
|
|
olympic
Senior Cruncher Joined: Jun 12, 2005 Post Count: 156 Status: Offline |
I experienced the same thing a few days ago, the task had been running for 30+ hours and was stuck at 53%. I aborted it but next time I'll restart BOINC and see what happens.
---------------------------------------- |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I got one that showed 3 days run-time; re-started BOINC and it went back to 1h48m run-time
 Looked at the results for this WU and there are already 18 returns  |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
For HPF2 there are 19 required, so if one of them comes back with invalid or error # 20 will be send, on and on until the 19 valid (which are all computed from a different seed), are complete.
----------------------------------------
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
keithhenry
Ace Cruncher Senile old farts of the world ....uh.....uh..... nevermind Joined: Nov 18, 2004 Post Count: 18665 Status: Offline Project Badges: 
|
It is important to check the reaults status on these before aborting. Restarting BOINC should not do any harm unless you're a long way from the last checkpoint. I've seen my recent HPF2 WUs taking much longer that typical lately. My current one is over 13 hours and shows 87% with almost two hours to go. Looking at results status shows times ranging from 8 to 32 hours on those that have completed it. However, I've found the percent complete and time to completion for HPF2 WUs to be a bit deceptive, not in the malicious sense, but due to how progress does not seem to be at a constant pace. I've seen the percent complete just as much as 8-9 percent all at once. The time to complete seems to hover about a given value and then suddenly drop by as much as 30-45 minutes. It does seem that if you catch these values when they change, they are pretty accurate in the end.
---------------------------------------- |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
So do you think I should kill these tasks?
Or should I try rebooting the PC?  -Brink |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Brink,
----------------------------------------if you checked that most all in the result status detail for those WU's were done in a semi standard of around 4 hours, you got a bad 'seed' (excuse the pun). Assume that credit was awarded for the first 11 like 30, than u crunching 2 more days each is not equitable..... send them to the here-after. Would though appreciate u send a mail to contact_us with device ID's and WU names, as 2 simultaneous with one member on 2 machines could have a common denominator. cheers
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
|