| 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: Tasks are not checkpointing proporly |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 54
|
|
| Author |
|
|
yose-ue
Cruncher Joined: Dec 27, 2008 Post Count: 21 Status: Offline Project Badges: 
|
Up until a couple of days ago I didn't have any problems running cep2 tasks. Now when I run them they will get up to about 25 to 30 percent then start over at about 2%. The last checkpoint was at 8 minutes even after the cpu time was at over 3 hours. I have aborted a task before because it was at over 20 clock hours and the cpu time was at less than 2 hours. I was running a childhood cancer project at the same time and the difference between the clock time and cpu time was only about a half hour. I reset the project after I aborted the previous project thinking that would fix the problem. I think I will have to abort the current project because it was at 25% with cpu and wall clock similar then I checked on it a half hour later and it was down to 4% complete with over 3 hours between cpu and wall clock but at least the checkpoint time changed to 25 minutes. Any suggestions?
cpu p4 at 3ghz running boinc 6.10.17 on ubuntu linux 2gig system memory |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
The unfortunate thing is that 6.10.17 is giving the impression things are fine, when the CPU time is stalling. Never that "at a glance" picture we had up through 6.2.28. Suggest to do a few TOP in a terminal window or start System Monitor to see if something else is eating time. Yesterday I found the sound-indicator eating 80% on a continuous basis after playing a little music of a internet radio site and had to boot to get normality to return.
----------------------------------------
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
yose-ue
Cruncher Joined: Dec 27, 2008 Post Count: 21 Status: Offline Project Badges:
|
I checked system monitor and the cep2 task is averaging about 97% of cpu
|
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Don't know. At the very least if not already done, plz follow recommendation for this project and set "Leave application in memory when suspended/preempted". This makes sure a task does not continuously regress when BOINC decides to pause for one reason or the other. The checkpoints on a P4 will be far apart. I've even seen 5 hours on my Q6600 between checkpoints, and that's true CPU time.
----------------------------------------
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I had 7 units that are back to zero even though I checked the left in memory and changed the option to change all the units ... minutes to 780 minutes ...
This did not prevent the two units that were calculated starting from scratch .... |
||
|
|
X-Files 27
Senior Cruncher Canada Joined: May 21, 2007 Post Count: 391 Status: Offline Project Badges: 
|
I can also concur on this issue. BM task property says it had checkpointed, so i restart, then viola back at zero. I believe if you restart without reaching about 25% first , "resume to checkpoint" does not work. But otherwise it will.
----------------------------------------  |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
A novel observation, this 25%.
----------------------------------------"change all the units ... minutes to 780" is probably the application switch time. Switching in general does not happen unless the first or subsequent checkpoint has been written **, which begs the question if anyone has been changing that value "Tasks checkpoint to disk at most every xxx seconds" aka Write to Disk (WTD). Mine are 300 seconds. The description is a bit ambiguous and should be read as that the WTD interval is minimum xxx seconds. The ones during the interval are skipped, not postponed. edit: ** except when the client is in high priority processing panic state.
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 1 times, last edit by Sekerob at Oct 4, 2010 6:16:05 PM] |
||
|
|
yose-ue
Cruncher Joined: Dec 27, 2008 Post Count: 21 Status: Offline Project Badges:
|
This was my original post. I have leave in memory checked. The tasks will run for 3 or 4 hours then start over run another 3 or 4 hours and start over. It is not starting over from zero it will checkpoint once or twice at the beginning. I didn't look before but another strange feature is if you have the screansaver running it will list progress at zero even if the progress in the task list may be as high as 40%. I didn't have any problems running cep2 before 9\28 after that day they just keep cycling. I am now running c4cw and hfcc and have no problems with those projects. The problem with the task repeatedly starting over is just with cep2. If I remember correctly there was an update to linux kernel around that time (my computer is set to automatically check for updates) but if that caused the problem then it should be affecting a lot of computers not just me.
|
||
|
|
I need a bath
Senior Cruncher USA Joined: Apr 12, 2007 Post Count: 347 Status: Offline Project Badges:
|
I have issues too. I have one task that started today 3 and a half hours ago and the cpu time is only 40 minutes. There is definitely something afoot. I'm gonna stop all until this gets figured out.
---------------------------------------- |
||
|
|
kateiacy
Veteran Cruncher USA Joined: Jan 23, 2010 Post Count: 1027 Status: Offline Project Badges:
|
Is there any chance that some change that was made in the Windows beta version of CEP2 was also ported to the Linux version? It seems as if a lot of us who've been crunching CEP2 for quite a while on Linux are now experiencing problems that we didn't used to have.
---------------------------------------- |
||
|
|
|