Index  | Recent Threads  | Unanswered Threads  | Who's Active  | Guidelines  | Search
 

Quick Go »
No member browsing this thread
Thread Status: Active
Total posts in this thread: 8
[ Jump to Last Post ]
Post new Thread
Author
Previous Thread This topic has been viewed 1032 times and has 7 replies Next Thread
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
confused Why such long checkpoints ?

Hello !

After "why such short deadlines", I was just wondering : why such long interval between checkpoints ?
It seems to me that in the last few weeks, the time between 2 checkpoints has been dramatically growing. Is it true or am I dreaming ?
Since I use my 2 computers only +- 2h and then stop and then +- 2h again, etc., a 30-40 min. interval between checkpoints seems enormous ! So it even happens sometimes, that when I restart my computer, I've got to crunch the same 30-40 min. ! What a waste of CPU !
I know that most of you are running 24/24h, but is it possible to reduce the interval to... +- 15-20min ?

Thanks.
Duanra.
[Sep 11, 2006 7:55:22 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
Re: Why such long checkpoints ?

See here:

http://www.worldcommunitygrid.org/forums/wcg/viewthread?thread=8553

Basically, we're encountering a bunch of work units where it is non-trivial to checkpoint more regularly.
[Sep 11, 2006 8:17:30 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
Re: Why such long checkpoints ?

Duanra,

The people who create the WUs are aware of the problem and they do everything they can to put in as many checkpoints as they can. However there is a point where adding more checkpoints wastes more time than it saves. So I'm afraid the answer to your question is "no".

When you stop using your computer after +-2 hrs, would it be possible for you to hibernate your computer rather than turn it off? If you can do that then the WU will resume exactly where it left off and you won't lose any time.
[Sep 11, 2006 8:56:30 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
Re: Why such long checkpoints ?

When you stop using your computer after +-2 hrs, would it be possible for you to hibernate your computer rather than turn it off? If you can do that then the WU will resume exactly where it left off and you won't lose any time.


I know but on BOTH my computers, each time I restart from "hibernating", standby mode or whatever the name is, I receive this message (+ see my thread in BOINC client support) :

11/09/2006 9:34:42|World Community Grid|Task faah0777_bdb863_mx1htf_dry_0B_0 exited with zero status but no 'finished' file
11/09/2006 9:34:42|World Community Grid|If this happens repeatedly you may need to reset the project.
11/09/2006 9:34:42||Rescheduling CPU: application exited

Off course, I tried to reset the project...always the same !
I've got this problem only with WCG and boincsimap WU's, not with Einstein@home or rosetta@home WU's.
The only difference with those projects is : the first ones don't have graphics ; the last ones do have graphics.
So I'm eagerly waiting for the graphics in faah WU's and I'll see if something changes...
Meanwhile......... the interval between checkpoints is still a bit too long for me (even if I know techs can't change anything for a moment).
Hopefully, I found that each checkpoint increase the disk space used by WCG. So I look sometimes to see if the application checkpointed or not.

Thanks for your support.
Duanra
----------------------------------------
[Edit 1 times, last edit by Former Member at Sep 11, 2006 9:29:32 AM]
[Sep 11, 2006 9:27:39 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Sekerob
Ace Cruncher
Joined: Jul 24, 2005
Post Count: 20043
Status: Offline
Reply to this Post  Reply with Quote 
Re: Why such long checkpoints ?

Duanra,

Though computers are hardware, they behave like human sometimes...quirky. Please refer to the DavidHobbs Hibernation manual. The immediate thought on hibernation (it has to be hibernation, not standby), is that possibly u have or have not the 'pre-empt' option on in your WCG BOINC profile. I've had trouble in the exact opposite of others, thus flipping the switch might work for u...... if Yes, set to No, if No, set to Yes.

ciao

PS, these are my settings:

Disk and Memory Usage
Write to disk at most every: 900 seconds
Leave applications in memory while preempted? YES
----------------------------------------
WCG Global & Research > Make Proposal Help: Start Here!
Please help to make the Forums an enjoyable experience for All!
[Sep 11, 2006 9:47:37 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
Re: Why such long checkpoints ?

I'm going to take a wild guess here, maybe i'm wrong, but it sounds like you think hibernate and standby are the same thing. They are not. Could that be the problem? Maybe you aren't hibernating it properly? When in hibernation mode you can unplug the computer and it will restart properly and resume. You can't do that in standby mode. It can take a minute or 2 to get into hibernate mode because it has to save a lot of info to the disk. If you don't wait long enough and unplug before it's properly hibernated then you'll have problems when it restarts.

From the messages it looks like some of BOINC's state files got corrupted. Perhaps you put it in standby and then unplugged it or perhaps it's a laptop and the battery went dead while in standby. Perhaps you did not wait long enough for it to hibernate before unplugging. All of those vould cause BOINC's state files to be corrupted.

Anyway, I think the best way for you to fix it is to uninstall BOINC, delete the BOINC directory and then reinstall BOINC from scratch. You could try to find the corrupted state file and fix it but that could take more time than a clean reinstall.

Then make sure you're hibernating the computer rather than just putting it in standby mode. I assure you this has nothing to do with graphics. And there are differences between the projects other than the graphics.

The increase in disk space used could be due to something other than the checkpoint. I am not sure that is a reliable test for checkpoint. For example, the occupied disk space could increase due to a download.
[Sep 11, 2006 10:28:11 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
Re: Why such long checkpoints ?

The difference between hibernation and standby - read http://www.worldcommunitygrid.org/forums/wcg/viewthread?thread=8583
[Sep 11, 2006 3:09:08 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
Re: Why such long checkpoints ?

Thanks everybody.
Unfortunately, I do know the difference between hibernation and standby mode. (In French, hibernating is called "prolonged stand by mode").
As for the "unplugging", I have a laptop (that's true) BUT I also have a PC which always stays plugged : so I fear this is not the cause of my problem.
However, I shall try to uninstall and reinstall all BOINC (and BOINC related ?) softwares.
If this still doesn't work, then I shall resume this thread in BOINC client support forum, which I think shall be more appropriate for this discussion.

Again, thank you.
Duanra
[Sep 11, 2006 4:50:16 PM]   Link   Report threatening or abusive post: please login first  Go to top 
[ Jump to Last Post ]
Post new Thread