I am having global problems with this program (see other posts: \'windows x64\' and \'calculating yearly data\'). Now the cpdn has downloaded and aborted several models on the trot. I have instructed the BOINC manager to prevent any new work. I have 1.78Gb of data in the cpdn folder, including some of the models downloaded. Please would someone help me tidy this installation up?

I am not sure that the program is keeping track of its bits and pieces. I have variously reached between 4% and 13% of several models before the program runs into trouble. I have some of these data in backup.

Please advise

---

Hi,

As far as I can tell the models have have been manually aborted. Does anyone other than you use the computer? If so, it may be worth explaining the difference between \'abort\' and \'suspend\'... !

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/result.php?resultid=5360946

If you have a backup of the model *before it was aborted* then restoring should work.

You can get rid of the space taken up by the aborted units just by deleting their folders.

The \'calculating yearly data\' message is harmless, and you can ignore it. It\'ll stick around for a full model year (i.e., between 12 hours and 36 hours depending on the speed of your PC).

---

I'm a volunteer and my views are my own.
News and Announcements and FAQ

Hi,

As far as I can tell the models have have been manually aborted. Does anyone other than you use the computer? If so, it may be worth explaining the difference between \'abort\' and \'suspend\'... !

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/result.php?resultid=5360946

If you have a backup of the model *before it was aborted* then restoring should work.

You can get rid of the space taken up by the aborted units just by deleting their folders.

The \'calculating yearly data\' message is harmless, and you can ignore it. It\'ll stick around for a full model year (i.e., between 12 hours and 36 hours depending on the speed of your PC).

************************

Hi,

Suspending the model, rather than aborting, simply downloaded another until the \'no new work\' was activated. The crashing models were manually aborted after downloads got stuck: repeatedly trying and failing to download the same 2 hadcm3***.exe files, while continuing to request more work (and, alas, getting it). Some of the models were crunched for 3-5 model years, but didn\'t write data either to disk or pagefile and had to restart at timepoint 0 after a crash. There are two kinds of instability: shutting down without loss of data, and shutting down with loss of all data.

At present, i am crunching the BBC models using the BBC version of BOINC as they have earlier deadlines This BOINC will only stay afloat if it is restricted to one processor, so there is a possibility that overheating is an issue, although the probes indicate against this. (The \'statistics\' page is also much easier to read than the cpdn: dark on light graphics wins hands down.)

I have several more models to crunch: at least 3 (although i am not sure how many are viable) from cpdn to be crunched by BOINC 5.2.13 (This version of BOINC is more stable than the BBC and runs for >48 hours reliably). Can i crunch any of the models on either BOINC, depending on which performs better?. Or should i just delete all the models and get fresh ones when the backlog has been cleared?

Sorry for the hassle that this new work / abort sequence has caused.

---

You can use any version of the boinc manager, but be careful about the installation directory.

As an example, for going from the bbc to standard:
* Make a backup of the BBC one,
* uninstall it,
* rename the BBC directory \'c:\\program files\\climate change experiment\' to \'c:\\program files\\boinc\',
* then download whichever standard version you want (5.2.19 and 5.4.9 are available from the main download site boinc.berkeley.edu).

---

I'm a volunteer and my views are my own.
News and Announcements and FAQ

You can use any version of the boinc manager, but be careful about the installation directory.

As an example, for going from the bbc to standard:
* Make a backup of the BBC one,
* uninstall it,
* rename the BBC directory \'c:\\program files\\climate change experiment\' to \'c:\\program files\\boinc\',
* then download whichever standard version you want (5.2.19 and 5.4.9 are available from the main download site boinc.berkeley.edu).

Not quite there: have three BBC models loaded and one running under BOINC 5.4.9. cpdn displayed as a project.

Condition satifactory

However, BOINC manager indicates the presence of cpdn as a project but there are no models loaded for this project, although there are 5 hadcm3 folders and a whole load of zipped files, most of which carry similar file names to the folders in C:\\Program Files (x86)\\boinc\\projects\\climateprediction.net. It is these files that i wish to tidy up, so that i know what to backup and what to discard.

Getting there slowly, thanks for the trouble you\'re taking

---

I started reading this thread because it said it was about exceeding daily quota. I have found no mention of this problem. I get a message from cpdn that I have exceeded my quota of 2 results how do I fix this.


Louie

There is no human condition so miserable that it can not be made worse by the presence of an officer of the law.

Mark Twain

The original poster\'s problem was explained without a need to explain about quotas.

The daily quota is to prevent people from a constant cycle of:
1) Download a model
2) Quickly crash the model
3) Start again at step 1

If you\'ve reached your limit, them you have a problem with your computer, either hardware, software, or operator useage.
As you have your computers hidden, no detailed advice can be given.

There are hints, tips, and advice, all linked from the READMEs here.
Hope they help.

---

Backups: Here

I have changed my prefrences so my computer should now show. I am a big dummy and did not find what I needed in the readme, of course I did not know what I was looking for.


Louie

There is no human condition so miserable that it can not be made worse by the presence of an officer of the law.

Mark Twain

Your computers don\'t yet show on your web page but I\'d recommend looking at the README about model crashes and problems, particularly item #5 by Mike. And find Les\'s README post about how to back up the BOINC folder. So if a future model crashes, in most cases you can restore the backup and continue it

---

Cpdn news

Louie
The servers have finally updated the user data, so we can see your model data.

You\'re getting the \"107nnnnn\" errors, so I\'m going to point you to this thread at the BOINC site.

For the 1073741502 errors, if you\'re NOT running SETI and also getting this error there, DON\'T report them there. Fixing things for the other \"107\'s\" may help somewhat.

There are two main errors you\'re getting:

-1073741502 (0xc0000142) (Cannot load DLL) : The most recent version (5.10.30) of the Boinc manager says that it has fixed these. See the \'News and Announcements\' thread. These errors happen under two circumstances that I\'m aware of : Firstly, if the PC is out of system resources, and just about to crash, you\'ll find that it won\'t let you open any new applications. When the Boinc manager tries to start a project\'s workunit, it will receive the C0000142 error. The second situation is during system shutdown. If Boinc doesn\'t realise that a system shutdown is taking place, it will try to start a new workunit, and again, this error will occur.

1073807364 (0x40010004) : I\'ve seen these ones when the graphics are displayed, and I try to shut the graphics window using the \'X\' in the upper right. So I\'d suggest not using the screensaver, and possibly also updating the system\'s graphics card driver, if there is an update available. From the description, it sounds like they can also happen at system shutdown.

So an additional suggestion is to right-click on the Boinc icon and \'exit\', just before shutting down Windows. This gives time for the Boinc manager + client to shutdown cleanly before Windows forces them to exit. You can also extend the shutdown time by following the instructions in the READMEs (link in my signature, see \'crashes and other problems\', \'Vista\', \'How to slow down shutdown\').

---

I'm a volunteer and my views are my own.
News and Announcements and FAQ