| 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 |
|
|
keithhenry
Ace Cruncher Senile old farts of the world ....uh.....uh..... nevermind Joined: Nov 18, 2004 Post Count: 18667 Status: Offline Project Badges: 
|
Well, Sek. This is a bit of a catch-22 situation. If you abort the task, you get no points. That's not bad if you can spot the problem early but if you're wrong, you delay that WU's completion and could get in trouble with the limits on returning error results. If you let it run, you may end up getting points, perhaps at the outlier level plus you've wasted a lot of crunching time. Also, when you're comparing your situation with those who've completed the WU in Results Status, you have no way of doing an accurate apples to apples comparasion. You know the other crunchers are running the same OS probably but you could a low end machine comparing yourself to high end machines that normally would complete the WU faster than you anyway. With HPF2 as well, with the Send 19 Quorum of 11, I'm seeing CPU times that vary by as much as a factor of five. By the time you've gotten to the point you can say, something really is wrong, it's way late. I would add though that if you're running a 5.8.x level of BOINC, you confirm that the WU *really is* using CPU time. I saw 5.8.8 tell me BOINC was crunching when task Manager showed the science app idle. In that case, shutting down BOINC and restarting it fixed that.
---------------------------------------- |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
One does not get in trouble over one or the other job cancellation. The closing of the feed is slower than the opening.... get 1 good job back after a bad and it's almost back to normal.
----------------------------------------One can compute for oneself if things are within parm. 1. Usually, on these overextended jobs, the quorum has already been established and credit determined.....HPF2 is usually withing 24-36 hours with the first 11 back. 2. Take your own time and multiply by hourly claim. If at 50% on CPU and project claim already multiples or awarded credit, consider the job to be bad.... that is, if the machine is known to claim within the margins of other crunchers. Why these wide varsities in claims exist i have a theory on. Why one or the other takes longer, we know.....some attempts take considerable longer than the other. Note... knreed advised that since about last Friday (Mar.2), the run times have been increased from 4 to 6.5 hours for a standard machine. For a device that used to take 8 on average, that translates to 13 hours.
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Dieter Matuschek
Advanced Cruncher Germany Joined: Aug 13, 2005 Post Count: 142 Status: Offline Project Badges:
|
Just for information:
----------------------------------------Today I've got the same problem: HPF2 WU stuck at 3.511 % for 16 hours. I didn't abort it but exited and restarted BOINC. Then all was quite normal. The WU reached 4% after some minutes and finished in some 9 hours.  (My guess is that it's a feature of the algorithm.)  Ask not what the world can do for you - ask what you can do for the world. [Edit 1 times, last edit by Dieter Matuschek at Apr 13, 2007 6:49:39 PM] |
||
|
|
E. Frijters
Senior Cruncher The Netherlands Joined: Apr 26, 2007 Post Count: 228 Status: Offline Project Badges: 
|
Until today I got WU's that had a "time to completion" of 07:20 hrs.
----------------------------------------Now I receive work that probably need 13:20 hrs of processing... I guess this is normal? 
Former grid.org slave
  |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
WCG try to design work units so that they take about 10 hours on the average computer. However, work units vary. I've heard of work units taking a week to complete.
But that's rare. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
74hours -> 49%
|
||
|
|
E. Frijters
Senior Cruncher The Netherlands Joined: Apr 26, 2007 Post Count: 228 Status: Offline Project Badges:
|
Got a new large one : 67 hours running, time to completion: 21 hours...
----------------------------------------[update:] Time "to completion" is increasing constantly... I hope I get extra points for crunching larger WU's.... 
Former grid.org slave
----------------------------------------[Edit 2 times, last edit by E. Frijters at May 21, 2007 8:32:33 AM] |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
LONG as it makes progress (for UD one can see 0.1% steps in the Graphics screen), you're fine. Add the CPU speed and, if inclined, someone will tell if you have to call Houston
---------------------------------------- The slowest machine took 204 hours to finish a HCMD yesterday 
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
E. Frijters
Senior Cruncher The Netherlands Joined: Apr 26, 2007 Post Count: 228 Status: Offline Project Badges:
|
LONG as it makes progress (for UD one can see 0.1% steps in the Graphics screen), you're fine. Add the CPU speed and, if inclined, someone will tell if you have to call Houston The slowest machine took 204 hours to finish a HCMD yesterday How long can 0,1% of progress possibly take? Mine is now at 70,1% for some 45 minutes... I'll leave the graphic screen on to check any progress.
Former grid.org slave
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
One thing you can do when you have a work unit going longer than you expect is check the work unit out in the Results Status link as per my example below.
It will show you all 19 work units and their current progress, ie either pending validation, in progress, error, inconclusive etc. with your personal work unit high lighted. This will give you a good idea who long the work unit should probably take and if you should consider stopping your unit. In my example below results show: 12 units pending validation - time range 18.95 (slow machine) to 4.8 (fast machine) 7 units In progress In this case I know I am waiting for 3 more units to complete before work unit is validated. Therefore depending on your computer for example if it is showing 40 hours and 57% complete you may want to stop your work unit as only 3 more units need completion before validation will start. I also find this allows me to keep track of how many of my work units have complete and validated and how many are pending. For example I now have 3 work units pending, 2 waiting on one more completion and the one below waiting on 3. I hope this helps others out how are experiencing long work units. World Community Grid Workunit Status Project Name: Human Proteome Folding - Phase 2 Created: 05/19/2007 07:32:30 Name: lc064_00043 Minimum Quorum: 15 Initial Replication: 19 Result Name Status Sent Time Time Due / Return Time CPU Time (hours) Claimed/ Granted BOINC Credit lc064_ 00043_ 7-- In Progress 05/20/2007 04:27:36 05/29/2007 04:27:36 0.00 0.0 / 0.0 lc064_ 00043_ 16-- Pending Validation 05/20/2007 04:05:13 05/21/2007 03:52:14 18.95 90.8 / 0.0 lc064_ 00043_ 13-- Pending Validation 05/20/2007 04:01:50 05/20/2007 20:47:33 14.08 54.5 / 0.0 lc064_ 00043_ 4-- In Progress 05/20/2007 03:59:50 05/29/2007 03:59:50 0.00 0.0 / 0.0 lc064_ 00043_ 0-- In Progress 05/20/2007 03:54:41 05/29/2007 03:54:41 0.00 0.0 / 0.0 lc064_ 00043_ 2-- In Progress 05/20/2007 03:34:37 05/29/2007 03:34:37 0.00 0.0 / 0.0 lc064_ 00043_ 1-- Pending Validation 05/20/2007 03:32:17 05/20/2007 11:51:03 7.16 56.7 / 0.0 lc064_ 00043_ 14-- Pending Validation 05/20/2007 03:24:34 05/21/2007 04:50:33 13.75 84.6 / 0.0 lc064_ 00043_ 17-- Pending Validation 05/20/2007 03:22:11 05/20/2007 19:21:01 13.42 85.4 / 0.0 lc064_ 00043_ 5-- Pending Validation 05/20/2007 03:18:30 05/21/2007 00:32:44 14.27 95.9 / 0.0 lc064_ 00043_ 3-- In Progress 05/20/2007 03:17:26 05/29/2007 03:17:26 0.00 0.0 / 0.0 lc064_ 00043_ 9-- Pending Validation 05/20/2007 03:09:36 05/20/2007 21:45:03 10.24 82.9 / 0.0 lc064_ 00043_ 12-- Pending Validation 05/20/2007 03:07:00 05/20/2007 18:47:11 6.10 71.4 / 0.0 lc064_ 00043_ 15-- In Progress 05/20/2007 03:04:39 05/29/2007 03:04:39 0.00 0.0 / 0.0 lc064_ 00043_ 11-- Pending Validation 05/20/2007 03:04:28 05/20/2007 11:57:42 5.65 63.3 / 0.0 lc064_ 00043_ 18-- Pending Validation 05/20/2007 02:59:47 05/20/2007 14:06:51 10.14 116.6 / 0.0 lc064_ 00043_ 6-- Pending Validation 05/20/2007 02:51:53 05/21/2007 00:42:45 13.57 88.7 / 0.0 lc064_ 00043_ 10-- Pending Validation 05/20/2007 02:51:51 05/20/2007 13:26:39 4.80 67.1 / 0.0 lc064_ 00043_ 8-- In Progress 05/20/2007 02:51:09 05/29/2007 02:51:09 0.00 0.0 / 0.0 close |
||
|
|
|