Questions and Answers :
Windows :
Crash (when uploading zip-file)
Message board moderation
Previous · 1 · 2
Author | Message |
---|---|
Send message Joined: 29 Sep 04 Posts: 2363 Credit: 14,611,758 RAC: 0 |
Yes, you\'ve done it right and it looks much cleaner now. Do the AMD records not merge? They appear identical so they should also merge. Cpdn news |
Send message Joined: 4 Sep 06 Posts: 79 Credit: 5,583,517 RAC: 0 |
Now they\'re merged too. But where is the old Duron I retired about a year ago? |
Send message Joined: 4 Sep 06 Posts: 79 Credit: 5,583,517 RAC: 0 |
Tomorrow one of the models will finish. I have borrowed an AMD to see if the troubled model can pass the \"point of trouble\". What should I do then? The finished model will appear again, the two others I will finish on the C2Q (now 30%, 23% on the folder that I am going to place on the AMD) so it will only be the troubled model I will run an the AMD. Should I abort them or suspend them while running the one model. (Hope you understood)\' Thank you Steinar |
Send message Joined: 4 Sep 06 Posts: 79 Credit: 5,583,517 RAC: 0 |
The crashing model went just fine passed the point of crash when I changed to an AMD and has run without problems for 24 hrs on the Intel after that:-) Good to see. I will post here when the model has finished in case someone would like to know. But most of all, good to see that a model that crashed 5 times at the same point with the Intel made it pass that point with the other platform. Only wish I knew this a bit earlier.. Steinar |
Send message Joined: 29 Sep 04 Posts: 2363 Credit: 14,611,758 RAC: 0 |
This is the model, apparently doing well, but you\'ll need to continue backing it up, just in case. About 8 years to complete. http://climateapps2.oucs.ox.ac.uk/cpdnboinc/result.php?resultid=7192227 PeteB discovered that he could get a continuously looping model past the loop point by moving it to the other sort of CPU. Not many people have a computer of each type to try this, but when it\'s been tried I think it\'s mostly been successful. I imagine that a loop is a type of calculation error, so making the model calculate differently often works. And where has your old Duron gone? I definitely saw it in your list of computers; that\'s why I wondered whether you could still use it to transfer the model to. First time I\'ve seen a computer that\'s run models disappear from a CPDN list. Check your name on BoincStats to see whether you\'ve somehow got an extra CPID and the Duron\'s being used by your doppelganger. Update all your projects on both computers even if you don\'t want new work. Cpdn news |
Send message Joined: 31 Dec 07 Posts: 1152 Credit: 22,133,755 RAC: 2,026 |
The crashing model went just fine passed the point of crash when I changed to an AMD and has run without problems for 24 hrs on the Intel after that:-) Good to see. I will post here when the model has finished in case someone would like to know. Hi, I am happy to hear that transferring the model from and intel processor machine to an AMD worked as well for you as it did for me (in the opposite direction). It can be heart breaking to loose a model after investing weeks or months in it. Keep making those backups. I make one every morning. That way you only loosed a few hours if there is a crash. |
Send message Joined: 4 Sep 06 Posts: 79 Credit: 5,583,517 RAC: 0 |
The model has been restored to AMD and passed the point and crashed again on the Intel. I\'m done with it. If someone will try to finish it, I can burn it and send it on a DVD. |
Send message Joined: 29 Sep 04 Posts: 2363 Credit: 14,611,758 RAC: 0 |
You did everything a person could to try to save this model. I think we have to conclude that something made the model itself unstable. Better luck with your next one. Cpdn news |
Send message Joined: 5 Sep 04 Posts: 7629 Credit: 24,240,330 RAC: 0 |
The project people WANT to know about models that don\'t reach their specified end point. This is part of what the project is about - to learn what tiny changes in the starting values do to a model, so that they can improve the accuracy of the climate modelling programs. Using every means possible to make a model reach the specified end point is just defeating this part of the project. |
Send message Joined: 4 Sep 06 Posts: 79 Credit: 5,583,517 RAC: 0 |
OK, see your point Les, but maybe the limit where what you state meets back-up is a little dodgy... Anyway, got a little hung up in this one since it was almost there. The model is terminated and I finish the others. Thank you for all help. Steinar |
©2024 climateprediction.net