http://climateapps2.oucs.ox.ac.uk/cpdnboinc/result.php?resultid=5618592

This model crashed on me right before the end. Where did I go wrong?

EDIT:
This was the error message relating to this model I guess

1/12/2007 2:26:37 PM|climateprediction.net|Unrecoverable error for result hadcm3lbm_amo1_25265204_0 (<file_xfer_error> <file_name>hadcm3lbm_amo1_25265204_0_16.zip</file_name> <error_code>-161</error_code></file_xfer_error>)

Or maybe it wasn\'t something you, or your box, did, but an error with the model itself. I\'ve heard that happens, too... Have any backups?

Model crashed: umshell1.f: P_TH_ADJ : NEGATIVE PRESSURE VALUE CREATED.

Three of my Models crashed with this error in the last month (at ~29/49/97%). Rerun failed at the same place with the same error. I don\'t consider yours a good candidate for recovery with a backup.

It\'s most likely the victim of its parameter mix. (The scientists also have to see what doesn\'t work -- even though it isn\'t a lot of consolation to see a failure that far into a Model.) Be assured that your work was not wasted; it will be used in the evaluations.

---

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

I hope so, it is a good many hours of computing. It would be a shame if it had been a waste of time. As for a backup, I\'ve made some but the last one is almost a month back (I blame the holidays <g>)

The \'negative pressure values created\' in the messages suggests that one or more of the initial parameter values was unviable. This can be the case even if the graph looks healthy, as yours did. The researchers also need to know about which parameter values \'work\' and which don\'t, so your computer time has in no way been wasted. The values for the parameters have clearly not been narrowly predetermined and must in some cases be at or beyond the limits of what seems possible or probable.

At a climate conference I attended in London a couple of years ago, this slightly adventurous aspect of the experiment\'s design was presented by one of the Oxford researchers as a good feature. No researcher from any of the other institutions represented spoke up to oppose this feature of our experiments.

I would let this model go and start a new one.

---

Cpdn news