| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
| World Community Grid Forums
|
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 27
|
|
| Author |
|
|
Loui_h20
Cruncher Joined: Aug 12, 2013 Post Count: 25 Status: Offline |
Windows has just finished updating itself and it wants to restart, but as you know hours of work will be flushed down the drain if it just went ahead with said restart.
----------------------------------------So in order to avoid that I've suspended 23 days of cached work units and will suspend wu's as they hit checkpoints. Right now I'm sitting here suspending any wu with a cp (check point ) of less than a hour of work. Does anyone know if there is a simpler way of doing this or will I have to sit here and keep an eye on boinc? Here we go just hit a 8hr cp!! Suspend!! ![]() |
||
|
|
jonnieb-uk
Ace Cruncher England Joined: Nov 30, 2011 Post Count: 6105 Status: Offline Project Badges:
|
Restarting immediately after a checkpoint is certainly a valid strategy when crunching CEP2. It's not relevant for FAAH since WUs are both significantly shorter and checkpoint more often. That just leaves MCM1. There was a time when restarting in the middle of crunching a WU was fairly certain to end with the WU going P/Ver when eventually completed. That was fixed though there have been a few recent reports that it may again be a problem.
----------------------------------------In the final analysis it's down to the individual cruncher how much time is spent micro-managing BOINC. Make sure you have LAIM checked before restarting. |
||
|
|
Loui_h20
Cruncher Joined: Aug 12, 2013 Post Count: 25 Status: Offline |
Thanks for the info.
----------------------------------------I've got one last WU to suspend after the cp. Hopefully things will be okay after the restart! At least I've learned how to handle this situation now and I hope it helps others out in the future from losing loads of time, effort and WU work. ![]() |
||
|
|
Barnsley_Tatts
Senior Cruncher Joined: Nov 3, 2005 Post Count: 291 Status: Offline Project Badges:
|
It's a shame there's no way to 'force' a check point.
----------------------------------------Throughout the WCG Community there must be many hours lost daily due to reboots, and closing down for the night etc. Maybe a force checkpoint function could be looked at some time in the future by the techs! ![]() |
||
|
|
Loui_h20
Cruncher Joined: Aug 12, 2013 Post Count: 25 Status: Offline |
How about a staggered check point shut down.
----------------------------------------Anyway! Here goes nothing or at least loads of hours of work! Restart time! ............................. Okay it seems that I lost ~1% from somewhere off the wu but resuming the wu's one after another seems to do the trick! Keeping an eye on bionctask messages at the same time to keep me informed of what bionc is getting up to. ............................. And don't forget to resume ALL cached tasks.....or you might leave your PC with nothing to munch on over the weekend [embarrassed]. ![]() [Edit 3 times, last edit by Loui_h20 at Feb 14, 2014 9:32:19 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello Barnsley_Tatts,
A forced checkpoint is not possible. A checkpoint can only be written when the project program reaches a point where a relatively small packet of data can store the whole state of the program. When giant work arrays are full of data in rapid change, saving is not reasonable. Lawrence |
||
|
|
cehunt
Senior Cruncher CANADA Joined: Oct 10, 2011 Post Count: 172 Status: Offline Project Badges:
|
Hi:
This morning when I started my PC, I found out I lost out on 15 hours of work on 3 WUs. Before I turned my laptop off last night, I noticed that I had 3 WUs that were almost finished but I thought when I checked off "leave applications in memory" setting in my profile, it would prevent the loss of my work. Apparently it did not. This is not efficient and discourages people from participating in projects like the Clean Energy Project. As of Feb 15, there is 104 days of work left so I will soldier on because I believe in the project. I do not need any badges in this project as I have accumulated almost 14 million points. Yes, 14 million points. Clive |
||
|
|
Falconet
Master Cruncher Portugal Joined: Mar 9, 2009 Post Count: 3315 Status: Offline Project Badges:
|
cehunt,
----------------------------------------For CEP2 that setting is only good when you hibernate or suspend? your computer. If you turn it off, you will lose all "uncheckpointed" progress. If you hibernate it won't, and hibernating is basically shutting down your computer while saving everything that was running. So, next time, hibernate instead. ![]() - AMD Ryzen 5 1600AF 6C/12T 3.2 GHz - 85W - AMD Ryzen 5 2500U 4C/8T 2.0 GHz - 28W - AMD Ryzen 7 7730U 8C/16T 3.0 GHz |
||
|
|
philperkins
Cruncher Joined: Nov 9, 2005 Post Count: 11 Status: Offline Project Badges:
|
I couldn't agree with you more about checkpoints. Due to the cost of running 2 X X5679 (350VA) I have had to cut back to using the off-peak tarrif which is just 7 hours in the night. I only recently discovered that unless you run the full task with no interruption, this is literally just a time-waster. 7 hours may achieve near 50% but even then MCM may flip to high priority, as that is another problem where it gets too many tasks hence suspending an already time-wasting CEP. It would appear that there is no fix?
It seems that now, in addition to having to be awake at silly hours to abort MCMs that will return "too late", I now have to stop doing CEP altogether. Please, someone, tell me that I am wrong. (but if I am, please be kind enough to reveal the secret of CEP with only a nightly 7 hour run) Very disappointed - Phil |
||
|
|
Falconet
Master Cruncher Portugal Joined: Mar 9, 2009 Post Count: 3315 Status: Offline Project Badges:
|
Phil,
----------------------------------------If you set your computers to hibernate once those 7 hours end and enable "Leave Applications In Memory" you should be able to complete CEP2 tasks. ![]() - AMD Ryzen 5 1600AF 6C/12T 3.2 GHz - 85W - AMD Ryzen 5 2500U 4C/8T 2.0 GHz - 28W - AMD Ryzen 7 7730U 8C/16T 3.0 GHz |
||
|
|
|