Well 2 days ago,I left the pc open (the project was done at about ~85%) and I went to bed.When I wake up in the morning,the task had begun again from the start.Is it natural?
And all the work that was done till that moment(~85%) is lost or is it saved and sent?
Thank you,
Alexander

1) I suspect that your model crashed, and that you have started on a new model.
I can\'t get to the database at present to check, but you can see for yourself here when it\'s accessable again. Just click on the number under Result ID to see the details for a model.

2) Data is returned to the project once each model year by way of a \'trickle\', with extra data being returned every 10 years in the form of a zip file.
This and other matters are discussed / explained in the 5 README files, which can be accessed through a link in my signature below.

---

Backups: Here

<core_client_version>5.8.15</core_client_version>
<![CDATA[
<message>
� ������� ��� ����������� ����� ��� ������. (0x16) - exit code 22 (0x16)

. . . (large cut)

[01/28/08 00:10:44] TRACE [3908]: Retrieved the required desktop
Not a JPEG file: starts with 0x01 0xda
CPDN NetCDF Err #2 - No such file or directory
CPDN NetCDF Err #2 - No such file or directory
CPDN NetCDF Err #2 - No such file or directory
CPDN NetCDF Err #2 - No such file or directory
MainError:	10:18:22 PM	No files match the supplied pattern.
MainError:	10:18:22 PM	No files match the supplied pattern.
CPDN process is not running, exiting, bRetVal = 1, checkPID=0, selfPID=252, iMonCtr=1
Model crash detected, will try to restart...
CPDN process is not running, exiting, bRetVal = 1, checkPID=0, selfPID=252, iMonCtr=1
Model crash detected, will try to restart...
CPDN process is not running, exiting, bRetVal = 1, checkPID=0, selfPID=252, iMonCtr=1
Model crash detected, will try to restart...
CPDN process is not running, exiting, bRetVal = 1, checkPID=0, selfPID=252, iMonCtr=1
Model crash detected, will try to restart...
CPDN process is not running, exiting, bRetVal = 1, checkPID=0, selfPID=252, iMonCtr=1
Model crash detected, will try to restart...
CPDN process is not running, exiting, bRetVal = 1, checkPID=0, selfPID=252, iMonCtr=1
Model crash detected, will try to restart...
Sorry, too many model crashes! :-(

</stderr_txt>

Don\'t know what to make of this. Does this Model run with other Projects? Are your Preferences set to leave in memory when suspended? Do work when you are active?

---

"We have met the enemy and he is us." -- Pogo
Greetings from coastal Washington state, the scenic US Pacific Northwest.

Don\'t know what to make of this. Does this Model run with other Projects? Are your Preferences set to leave in memory when suspended? Do work when you are active?

No,the model was running alone.
And the preferences work all the time,active or not

So the model crashed(too pity,I was that close) but at least the results were not wasted.
Anyway I think 2100 hours(about that) are a lot for a project from a domestic pc and I do not intend to start from the beginning.I am going to delete ClimatePrediction and put another program.
I thank you everyone for your time.
Keep running Boinc :)

Alexander,

There are shorter models if you prefer (approximately 400 hours). Select HADSM3 in your account preferences and deselect HADCM3 and HADAM3.

Whatever you decide to do, thanks for your efforts.

Iain

Alexander,

There are shorter models if you prefer (approximately 400 hours). Select HADSM3 in your account preferences and deselect HADCM3 and HADAM3.

Whatever you decide to do, thanks for your efforts.

Iain

How can I do that?

If you select the \'Your account\' menu item, which is on left of this Web page. Then click \'View or edit\' for the \'climateprediction.net\' preferences: that will show the preferences page. Click on \'Edit climateprediction.net preferences\' towards the bottom of that page and then tick the options you want and clear the tick on the options you don\'t. For the short models, select \'UK Met Office HADSM3\'. The HADAM3 models are also quite short to run, but they needs lots of memory.

This link should get you there: http://climateapps2.oucs.ox.ac.uk/cpdnboinc/prefs.php?subset=project.

It also helps to take regular backups (about once a week) of your BOINC folder (while BOINC Manager is not running). A failure such as yours can then easily be recovered from the backup. I\'ve saved many WUs that way.

Hi, um ok i seem to have a similar problem so i thought i\'d post on here instead of making a new thread.

Ive been running the SM3 model, along with the AM3 and several other projects, however the SM3 model was at about 97% after around 200 hours cpu time then the next time i looked at boinc the progress has started again and is now on around 10% at 210 hours.
the last time my computer trickled anything was feb 25th just after the model entered into the 3rd phase, however since then there have been no trickles and no new claimed credit.

What i would like to know is, are the results from any continued work on this model going to be valid? or has the model crashed?
If the model is still producing usable results then i\'ll carry on running it.

any help on this matter would be very much appreciated.

Thanx, Alex


Latest Trickles Received
Time Sent (UTC) Host ID Result ID Result Name Phase Timestep CPU Time (sec) Average (sec/TS)
25 Feb 2008 08:35:43 830984 7224247 hadsm3fub_01dg_005928033_2 3 32,406 638,507 1.1590
24 Feb 2008 23:53:42 830984 7224247 hadsm3fub_01dg_005928033_2 3 21,604 625,476 1.1581
24 Feb 2008 18:14:45 830984 7224247 hadsm3fub_01dg_005928033_2 3 10,802 612,858 1.1579
23 Feb 2008 15:23:00 830984 7224247 hadsm3fub_01dg_005928033_2 2 259,248 599,571 1.1564

There are messages in the Messages tab of the BOINC gui which will tell you if a trickle was uploaded successfully.

These should be in sets of 4 lines; first BOINC will request the attention of a schedular, and then you get a line containing \"succeeded\", or several lines about a failure. (In which case the trickle_up_.... file will remain on your computer.)

... going to be valid ...

Only the researchers will know this.
But all data will be kept for analysis, so in that sense, it\'s recommended that you continue.

Sometimes models DO rewind to the start when something goes wrong. (I think that there\'s a post in the README files about it. See my sig below for a link.)
But there are also times when the person hasn\'t looked closely at the model\'s name, and it turns out that the model they\'re talking about has crashed, and what they\'re looking at is a new model.

As the servers are still trying to catch up with the workload after the recent shut down, it\'ll be a while before anyone can look at your results page to see if it\'s crashed, so you\'re the only one who can see anything on your computer.

Have a look at the graphics, and see what year the model is up to. If it seems to be OK, then you have two choices:
1) Wait while the model is recalulated all the way to the end again, or
2) Abort it and start a new model.

And try to remember what happen with and to your computer around the time that the model \"rewound\", and then try to prevent this happening again.
(This may be more obvious if there is an error message on the servers, when the info is available again.)

---

Backups: Here

Hey thanx for the speedy reply.

I checked the messages tab and every few hours cpdn is sending scheduler requests to send trickle up messages, it requests 0 seconds of work and 0 new tasks. A few seconds later it says scheduler request succeeded: got 0 new tasks. I looked on CP.net on the My Account section and now its displaying the trickle results. i heard that there were some problems with the servers or something so i guess that probably explains why there were no results between the 25th and today.

Also I checked the graphics and the model seems ok, like i said the years have started again from 1810 or whatever it was, its on 1817 now and it all seems alright.

it just seemed a little odd that the progress reset to 0% while the cpu time stayed the same, its definatly the same model from what i can tell.

its all good tho, im guna continue running it, im quite intrigued at whats going to happen in the 3rd phase (ive never got a model this far before)

anyway thank you again for your time. nice one!!

Alex

Several Phase 3 Trickles were logged yesterday, 29Feb; all seems to on track.

---

"We have met the enemy and he is us." -- Pogo
Greetings from coastal Washington state, the scenic US Pacific Northwest.

all the projects have reset but getting the message: access to reference site succeeded - project servers may be down

---

There\'s been a partial CPDN server outage (see the CPDN News thread, top of the forum Number Crunching section). The servers were supposed to be back to normal at about 9am UTC today ie 8 hours ago, but Milo may have taken longer than expected to get them all started again.

You can select \'Subscribe to this thread\' at the top of the News thread and receive an email notification whenever something new is announced.

---

Cpdn news