After Restart BOINC begins with 0 %

Message boards : Number crunching : After Restart BOINC begins with 0 %

To post messages, you must log in.

AuthorMessage
luQ

Send message
Joined: 20 Apr 07
Posts: 1
Credit: 1,465
RAC: 0
Message 39800 - Posted: 24 Apr 2007, 11:07:53 UTC

Hello,

me and nearly all people i know from the ESL Team have the same problem:

If the PC get restarted Boinc begins at 0 % with calculating. Even when the result reached 90 %, after a restart it becomes 0 %

Sorry for my bad english.

Can anybody help me?
ID: 39800 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Tom Philippart
Avatar

Send message
Joined: 29 May 06
Posts: 183
Credit: 834,667
RAC: 0
Message 39801 - Posted: 24 Apr 2007, 11:12:10 UTC - in response to Message 39800.  

Did you check the runtime? If it's still as it was before restarting, everything is ok, the percentage will catch back.

If the runtime is 0min 0 sec, there's something wrong, please check this.
ID: 39801 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
r0b

Send message
Joined: 18 Mar 07
Posts: 3
Credit: 27,955
RAC: 0
Message 39802 - Posted: 24 Apr 2007, 11:44:55 UTC

I have the same prob and elapsed time / time remaining get resetted, too.
ID: 39802 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile dcdc

Send message
Joined: 3 Nov 05
Posts: 1829
Credit: 116,388,135
RAC: 71,438
Message 39804 - Posted: 24 Apr 2007, 13:47:12 UTC

this is a known problem with Rosetta 5.59. The checkpointing is working ok apparently, so it's only the % that's restarting, and as Tom posted, and it will catch back up to where it should be if you let it.

However, if the CPU time is also restarting then work is being lost. Of course, work done since the last checkpoint will always be lost following a restart.

HTH
Danny

P.S. I hope ESL can keep up the good work!
ID: 39804 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Greg_BE
Avatar

Send message
Joined: 30 May 06
Posts: 5664
Credit: 5,711,666
RAC: 1,135
Message 39840 - Posted: 24 Apr 2007, 22:03:26 UTC - in response to Message 39804.  

just out of curiosity, would it be possible to include a save button to create a check point manually before restarting either ROH or windows?

this is a known problem with Rosetta 5.59. The checkpointing is working ok apparently, so it's only the % that's restarting, and as Tom posted, and it will catch back up to where it should be if you let it.

However, if the CPU time is also restarting then work is being lost. Of course, work done since the last checkpoint will always be lost following a restart.

HTH
Danny

P.S. I hope ESL can keep up the good work!


ID: 39840 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Mod.Sense
Volunteer moderator

Send message
Joined: 22 Aug 06
Posts: 4018
Credit: 0
RAC: 0
Message 39854 - Posted: 25 Apr 2007, 3:32:06 UTC - in response to Message 39840.  

...would it be possible to include a save button to create a check point manually?


Nope. But they are talking about checkpointing about every 5 minutes, if your preferences allow BOINC to write to disk that often. And so that would mean on average you'd only lose 150 seconds. So, hopefully that's even BETTER then the manual system you suggest, because it would be happening for everyone, not just those that know to perform a manual step first.

Rosetta Moderator: Mod.Sense
ID: 39854 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Tom Philippart
Avatar

Send message
Joined: 29 May 06
Posts: 183
Credit: 834,667
RAC: 0
Message 39859 - Posted: 25 Apr 2007, 10:18:46 UTC - in response to Message 39854.  

since ESL members often tune or overclock, is it possible that both of you tuned the shutdown of windows to accelerate it? It could cause the WU to be lost or resetted, try to close Boinc manually before restarting.
ID: 39859 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
r0b

Send message
Joined: 18 Mar 07
Posts: 3
Credit: 27,955
RAC: 0
Message 39861 - Posted: 25 Apr 2007, 11:33:30 UTC
Last modified: 25 Apr 2007, 11:33:58 UTC

thx for the hint, I actually use XP-Antispy and there is an option called "Quick Shutdown"

I'll test it later, whether it helps turning this option off or not
ID: 39861 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Greg_BE
Avatar

Send message
Joined: 30 May 06
Posts: 5664
Credit: 5,711,666
RAC: 1,135
Message 39888 - Posted: 26 Apr 2007, 0:04:14 UTC - in response to Message 39854.  

...would it be possible to include a save button to create a check point manually?


Nope. But they are talking about checkpointing about every 5 minutes, if your preferences allow BOINC to write to disk that often. And so that would mean on average you'd only lose 150 seconds. So, hopefully that's even BETTER then the manual system you suggest, because it would be happening for everyone, not just those that know to perform a manual step first.


So the write to disk every X minutes is the saving of a data checkpoint to the drive? So by changing your settings to write to disk every 60 seconds for instance is doing what?
ID: 39888 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Mod.Sense
Volunteer moderator

Send message
Joined: 22 Aug 06
Posts: 4018
Credit: 0
RAC: 0
Message 39892 - Posted: 26 Apr 2007, 3:09:41 UTC

As with many things BOINC, it is a control over how much BOINC uses your machine. So, if projects are honoring the settings, it allows you to LIMIT the amount of disk access performed (not counting the swap space). However, it doesn't work the other way around. I mean setting a low value doesn't CAUSE disk writes, or checkpoints to occur.

My understanding is that the Project Team is working on changes that will make checkpointing about every 5 minutes possible. If your preference were set to write at most every 20 minutes, then you'd be getting a checkpoint roughly every 20 minutes. If your preference were set to write to disk at most every 1 minute, you'd still only get a checkpoint when Rosetta is able to do so, about every 5 minutes.
Rosetta Moderator: Mod.Sense
ID: 39892 · Rating: -1 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile opernal

Send message
Joined: 26 Apr 07
Posts: 20
Credit: 4,598,641
RAC: 0
Message 39979 - Posted: 28 Apr 2007, 3:19:08 UTC

I'v got the same problem.
All the processing task return to 0% which have been calculated for hours after system restarted. What a bug!!
ID: 39979 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile David E K
Volunteer moderator
Project administrator
Project developer
Project scientist

Send message
Joined: 1 Jul 05
Posts: 1018
Credit: 4,334,829
RAC: 0
Message 39980 - Posted: 28 Apr 2007, 3:39:29 UTC

It's just for the display and doesn't actually mean work was lost. The cpu time is the best indicator of work done. This minor but seemingly major bug has been fixed in the next release that is currently being tested on ralph.
ID: 39980 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
hrsn___

Send message
Joined: 26 Mar 07
Posts: 5
Credit: 14,474
RAC: 0
Message 40042 - Posted: 29 Apr 2007, 12:29:46 UTC - in response to Message 39980.  

fine. good 2 know :)

OT question: are there any beta versions of boinc? ;)

It's just for the display and doesn't actually mean work was lost. The cpu time is the best indicator of work done. This minor but seemingly major bug has been fixed in the next release that is currently being tested on ralph.

ID: 40042 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Tom Philippart
Avatar

Send message
Joined: 29 May 06
Posts: 183
Credit: 834,667
RAC: 0
Message 40050 - Posted: 29 Apr 2007, 15:57:35 UTC - in response to Message 40042.  

fine. good 2 know :)

OT question: are there any beta versions of boinc? ;)

It's just for the display and doesn't actually mean work was lost. The cpu time is the best indicator of work done. This minor but seemingly major bug has been fixed in the next release that is currently being tested on ralph.


yes, you can download them here:
http://boinc.berkeley.edu/download_all.php

I tested the most recent ones, they are very buggy, I wouldn't recommend anyone to use them for real production.
ID: 40050 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote

Message boards : Number crunching : After Restart BOINC begins with 0 %



©2024 University of Washington
https://www.bakerlab.org