Climate Predication - Phase 2 Collapse

[UBT - John]

Just back form three very hot weeks in Australia - made the usual indent to the supplies of Carlton Cold. Swicthed on the computer to find that my Climate Predication model was AWOL - late in phase two - again! This is the second time this has happened  - again after a long haul trip -  and a lot of computer time has been wasted.
Any idea why this happens and is it recoverable from my Programme Files?
Meanwhile the computer has downloaded a new model and is happily chewing away to the exclusion of the other three projects.

[MikeMarsUK]

There are a list of FAQ items and so forth in the following location:

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


If you provide a link to the result on the website I can have a look at it and try to make some suggestions.  

But one suggestion I can make without looking at anything is it's well worth keeping a backup of Boinc when you're running the climate project.  There are several different ways to do this in one of the READMEs above.  It takes about 5 minutes in total, and if a crash happens you can go back and restore the backup.  I do it weekly, but some people do it daily.

When you say 'late in phase 2', does that mean you were running a 5-phase Sulphur model?  The coupled models always stay in phase 1 (although they do have hindcast/forecast sections, 1920-2000 + 2000-2080).

[UBT - John]

Mike,

The previous models were the 1 to 5 phase models that take the best part of 4000hrs to complete and I've lost them after about a 1000hrs. On each occasion it was after a boot up of the computer following a long haul trip.

Thanks for the advice - I'll follow the link and do a backup.  

Sorry for the late reply but I've been away...

Cheers,

John.

[MikeMarsUK]

The current models (coupled model) take perhaps 2.3x the time of the sulphur model, the coupled models were released late Febuary '06.