I\'m having this problem on my XP machine. Whenever I let my teenage daughter use IMVU, CPDN usually crashes within about 15 minutes. The following messages are typical:

24/04/2007 19:29:13|climateprediction.net|Unrecoverable error for result hadcm3inct_cl7x_1920_160_05872841_1 (Incorrect function. (0x1) - exit code 1 (0x1))
24/04/2007 19:29:13|climateprediction.net|Deferring scheduler requests for 1 minutes and 0 seconds
24/04/2007 19:29:16||Rescheduling CPU: application exited
24/04/2007 19:29:16|climateprediction.net|Computation for task hadcm3inct_cl7x_1920_160_05872841_1 finished

I now get her to \"snooze\" BOINC while she\'s using it, and since then it hasn\'t crashed (touch wood).

Does anybody have any idea why this is happening please?

---

Brian

Error 1 is mostly associated with shutting down too quickly for the modelling program to close all of it\'s files.
This may be because Windows is closed without first Exiting BOINC, or becuause the power is simply turned off with the program running.
And Vista is worse, because fast turn off is one of it\'s touted features.

The only guaranteed way to save the model is:
(Menu) Suspend BOINC,
Wait until it SAYS \"Suspended\", then
(Menu) Exit from BOINC.

Anything less is \"Russian Roulette\".

---

Backups: Here

Error 1 is mostly associated with shutting down too quickly for the modelling program to close all of it\'s files.
This may be because Windows is closed without first Exiting BOINC, or becuause the power is simply turned off with the program running.
And Vista is worse, because fast turn off is one of it\'s touted features.

The only guaranteed way to save the model is:
(Menu) Suspend BOINC,
Wait until it SAYS \"Suspended\", then
(Menu) Exit from BOINC.

Anything less is \"Russian Roulette\".

Thanks for the suggestion, but Windows is not being shut down in this case. When restarting Windows (I never shut it down - it runs 24/7) I had never thought of closing BOINC first, and this has never caused me any problems - so far.

---

Brian

Boinc relies on the graphics drivers, and sometimes certain programmes which use graphics can cause it to crash. Usually it\'s games or TV cards, but perhaps this IMVU programme has a similar effect. The odd thing is that it\'s not an error code usually associated with a graphics crash.

Ask her to shut down Boinc before running IMVU - all she needs to do is to right-click on the icon, and select \'exit\'.

The other thing which may help is taking a backup at intervals - so that if one day she forgets, then you have something you can go back to. For instructions on backups, take a look at the other forum\'s \'news and announcements/README - backup and restore\' thread.

http://www.climateprediction.net/board/viewforum.php?f=36

---

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

Just as an aside regarding not shutting down BOINC:
Do you leave newly created text/Word files etc open when you\'ve finished writing them?
Do you leaves games running after \"you\'re\" finished playing them?
I hope not. And BOINC and it\'s science apps are no different.

*****

The other possibility for Error 1 is a display (graphics) problem.

Not having heard of IMVU before, I Googled it. This is from the WIKKI entry:
\"IMVU is a graphical instant messaging client....\"

So it looks like there is a display_memory conflict between IMVU and the climate model.

This is from one of your models that failed with Error 1:

CPDN process is not running, exiting, bRetVal = 1, checkPID=0, selfPID=13812, iMonCtr=1
Model crash detected, will try to restart...

It looks like the climate program ran out of resources and stopped working for long enough for BOINC to detect the loss of the science app, and deciding that it had failed, thus triggering the process of deleting the \"failed\" model, telling the server it had failed, and requesting more work.

JM7 (I think) mentioned something about a crash happening after 15 minutes of CPU starvation, 15 minutes being the time when the model\'s wrapper decides it\'s time to start another copy.

---

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

For error code 1, I\'m pretty sure that the boinc Wiki suggests that updating the graphics card driver MAY help. So it can be similar to a 107 crash. For how to do the update see the README about avoiding crashes, item #6 by Thyme Lawn.

The update will be good for the computer anyway. But I\'d still suspend the model while your daughter\'s on the computer. Mike suggests exiting from boinc as well.

---

Cpdn news

Based on this : http://www.radgametools.com/pixomain.htm

They write that it is a competitor of DX7 and OpenGL.

...
But I\'d still suspend the model while your daughter\'s on the computer. Mike suggests exiting from boinc as well.

That was mainly because of the worries about a graphics clash, but if it was due to CPU starvation then suspending would do just as well.

---

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

Many thanks for all the suggestions (while I was in bed). From what people have said, it looks more likely that BOINC is being starved of CPU time than it being a graphics issue. I\'ve just noticed that I have been losing SETI work units in a similar way.

I\'ll make sure in future that she always shuts down BOINC.

---

Brian

Error 1 is mostly associated with shutting down too quickly for the modelling program to close all of it\'s files.
This may be because Windows is closed without first Exiting BOINC, or becuause the power is simply turned off with the program running.
And Vista is worse, because fast turn off is one of it\'s touted features.

The only guaranteed way to save the model is:
(Menu) Suspend BOINC,
Wait until it SAYS \"Suspended\", then
(Menu) Exit from BOINC.

Anything less is \"Russian Roulette\".

I have just had the same problem. My computer froze and I had to turn it off, and when I turnes it on again, one of my two models had crashed.
The first model I ran last year also crashed, this time because I played an internet radio

..And today BOINC froze totally. I restarted the PC, reinstalled BOINC in the same folder from a back-up, but it kept on freezing. At last I had to abort the two models and I uninstalled BOINC, deleted the old folders and reinstalled everything.
When I downloaded BOINC again it was ver. 5.8.16, the one I uninstalled was 5.8.15. Any problems with that version?

The only thing I\'m currently aware of which makes Boinc freeze is the network connection - if it tries to make a request, but the networking layer accepts the request but doesn\'t return a result, Boinc will freeze for a period of time.

This is an issue with Boinc 5.8.*, and was OK in prior versions (5.4.* etc). If you run boinc with \'network activity suspended\' for most of the time, that should sort it out (only a sensible suggestion if you only run long-duration work units such as CPDN). If you run shorter projects too, then my advice would be to downgrade to 5.4.x.

Your climate models were probably OK.

If you still have the logs (stdoutdae.txt, etc), you may find lines such as \'communication deferred 1m:00s), and possibly also \'exit status zero but no finished file\'.

http://boinc.berkeley.edu/trac/ticket/113

http://boinc.berkeley.edu/trac/ticket/171

---

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