 |
|
Message boards : Number crunching : HadCM3 in a Loop
Message board moderation
| Author | Message |
|---|---|
 old_user456571Send message Joined: 11 Jun 07 Posts: 5 Credit: 235,100 RAC: 0 |
I\'ve been running the project for several months and its progressed normally until now. It reaches 2038, I think, and then goes back to 12/2029, and keeps on doing this. Is the program stuck - do I just let it keep goiing or what? Any help out there? Brian Collier |
MikeMarsUK Send message Joined: 13 Jan 06 Posts: 1498 Credit: 15,613,038 RAC: 0 |
Usually a coupled model will only loop 3 times (day, then month, then year), and then automatically abort itself. Is it this model, which last trickled in 2028 on the 17th Dec? http://climateapps2.oucs.ox.ac.uk/cpdnboinc/result.php?resultid=6605646 How much CPU time has it received since the 17th? I'm a volunteer and my views are my own. News and Announcements and FAQ |
|
old_user456571 Send message Joined: 11 Jun 07 Posts: 5 Credit: 235,100 RAC: 0 |
Hello Mike, The model keeps asking me for an internet connection which I provide - I\'m on dialup. How do I find out how much CPU time it has received since any date? My CPU is running at 100%. I have some more data which may be of help, this is a couple experiment and one is further ahead than the other, the stats are: a. hadcm3int_ckgr_1920_160_55866727_1 is at 75.03% and time to completion is reducing, but b. hadcm3int_ckgw_1920_160_55866732_2 is at 68.119% and time to completion is increasing! The graphics go back to 01/12/2029 and have progressed as far as 2040 this morning before going back to 12/2029. any other data I can give you? Cheers Brian |
|
MikeMarsUK Send message Joined: 13 Jan 06 Posts: 1498 Credit: 15,613,038 RAC: 0 |
... How many hours are displayed on the CPU time so far column in the Boinc manager? The CPU time at the 17th Dec trickle is 6,889,761 seconds, which is 1,913 hours. From what I can see on the website, you have two models on that PC at the moment - one is up to 2040, and the other is the suspect one. I find it odd that a model would be going back to 2029 from 2040, never heard of that behaviour before. I'm a volunteer and my views are my own. News and Announcements and FAQ |
|
old_user456571 Send message Joined: 11 Jun 07 Posts: 5 Credit: 235,100 RAC: 0 |
... CPU time for _2 is 1926.21.44 and for _1 is 2118.20.31 |
|
old_user456571 Send message Joined: 11 Jun 07 Posts: 5 Credit: 235,100 RAC: 0 |
... Mike, I\'m replying again to your message. CPU time is as follows: hadcm3inct_1920_160_55866732_2 is 1926:20:09 hadcm3inct_ckgr_1920_160_55866727_1 is 2131:54:07 I suspect that the model isn\'t in a loop but that _1 is working fine and the problem is with _2 which is hanging up and the graphics are alternating between the two models or whatever they are. I need to know whether it is worth continuing with this experiment or if it has gone irretrievably wrong. |
|
MikeMarsUK Send message Joined: 13 Jan 06 Posts: 1498 Credit: 15,613,038 RAC: 0 |
If the model keeps \'losing\' CPU time somehow, then I\'d recommend aborting it. Is there anything in the Boinc messages log which indicates what happens when the CPU time goes? What does \'Task Manager\' / \'Processes\' show, are there two tasks getting CPU time or is there only one? I'm a volunteer and my views are my own. News and Announcements and FAQ |
|
old_user456571 Send message Joined: 11 Jun 07 Posts: 5 Credit: 235,100 RAC: 0 |
Hi again, Herewioth the message log, I can\'t see anything in it which indicates what is going on - 05/01/2008 13:00:37||Starting BOINC client version 5.8.16 for windows_intelx86 05/01/2008 13:00:37||log flags: task, file_xfer, sched_ops 05/01/2008 13:00:37||Libraries: libcurl/7.16.0 OpenSSL/0.9.8a zlib/1.2.3 05/01/2008 13:00:37||Data directory: C:\\Program Files\\BOINC 05/01/2008 13:00:37||Processor: 2 GenuineIntel Intel(R) Pentium(R) D CPU 2.80GHz [x86 Family 15 Model 4 Stepping 7] [fpu tsc pae nx sse sse2 mmx] 05/01/2008 13:00:37||Memory: 3.00 GB physical, 4.34 GB virtual 05/01/2008 13:00:37||Disk: 142.08 GB total, 69.15 GB free 05/01/2008 13:00:37|climateprediction.net|URL: http://climateprediction.net/; Computer ID: 693623; location: (none); project prefs: default 05/01/2008 13:00:37||No general preferences found - using BOINC defaults 05/01/2008 13:00:38|climateprediction.net|Restarting task hadcm3inct_ckgw_1920_160_55866732_2 using hadcm3i version 540 05/01/2008 13:00:39|climateprediction.net|Restarting task hadcm3inct_ckgr_1920_160_55866727_1 using hadcm3i version 540 There are still two taks being shown in the Task Manager the has had a CPU time of 2148:29:13 and the other 1926:33:17. The simple view continues to alternate between the two views showing ...732 at 01/12/2029 with the clock cucling very slowly. I have a feeling that something has gone wrong and I should abort. Can you recommend what i should run in its place or will that automatically be presented to me? |
|
Send message Joined: 5 Sep 04 Posts: 7629 Credit: 24,240,330 RAC: 0 |
Can you recommend what i should run in its place or will that automatically be presented to me? You can choose the type of model from the climateprediction.net preferences on your Account page, but you need to do this BEFORE Aborting your current model. If you don\'t select one, then the server allocates a type at random. A new model will be automatically sent when BOINC realises that one of the processors doesn\'t have work and asks for more. |
|
MikeMarsUK Send message Joined: 13 Jan 06 Posts: 1498 Credit: 15,613,038 RAC: 0 |
If you look in the \'running your model\' README, there are a couple of links to interesting posts describing the different types of model available. (Link to READMEs in my signature). I'm a volunteer and my views are my own. News and Announcements and FAQ |
©2024 cpdn.org
