climateprediction.net home page
ClimateChange model same as BBC?

ClimateChange model same as BBC?

Questions and Answers : Windows : ClimateChange model same as BBC?
Message board moderation

To post messages, you must log in.

AuthorMessage
old_user283420

Send message
Joined: 18 Feb 06
Posts: 6
Credit: 12,872
RAC: 0
Message 26493 - Posted: 1 Feb 2007, 17:42:53 UTC

Am I right in assuming that the climatechange model hadcm3 5.15 is the same as the BBC one hadcm3 5.15 (160 years).
I\'m asking because I think I have a bug in BionicView and before I post there I just wanted to check.
ID: 26493 · Report as offensive     Reply Quote
Profile Pooh Bear 27
Avatar

Send message
Joined: 5 Feb 05
Posts: 465
Credit: 1,914,189
RAC: 0
Message 26497 - Posted: 1 Feb 2007, 18:23:33 UTC

Correct.
ID: 26497 · Report as offensive     Reply Quote
old_user283420

Send message
Joined: 18 Feb 06
Posts: 6
Credit: 12,872
RAC: 0
Message 26530 - Posted: 3 Feb 2007, 1:06:59 UTC

Just checkng again because I still might have a problem with the time taken to complete a model.
I had until recently 2 BBC models running on a dual P3 1.2Ghz rig.
1 model crashed and was replaced with 1 from CPDN
hadcm3 5.15 hadcm3pbb_cc etc... which according to Boincview will take 135 days to complete @ 6.6s/TS

Yet the other model
hadcm3ln_0brw etc... is estimated to take 311 days @ 6.8 s/TS.

BOINC manager shows
CPU1 BBC model cputime 5352 completion 2115 = 7467
CPU2 CPDN cputime 79 completion 3182 = 3261
Is the CPDN model smaller or am I missing the point?


ID: 26530 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 26532 - Posted: 3 Feb 2007, 1:50:21 UTC

The models are the same, within the limits of different amounts of parameters taking a different amount of time to calculate.
For instance, a reference model, with no parameter forcings, will run a bit more quickly.

Because BOINC is optimised for the much shorter times of the other projects, the times that it displays for this project will only \'come good\' towards the very end of the model.
There\'s only one way to get an accurate estimate of the number of hours each will take, and that is to time how long each one takes between trickes, which is once a year, and mutiply this by the number of trickles remaining.
Then apply this to the average number of hours the computer is running each day, less the amount of time that YOU are using it. (Because BOINC won\'t be running while you\'re using it. At least, on one of the processors.)

My BBC model is showing 1031 hours done, and 1826 hours to go, whereas the cpdn model is showing 1017 done, and 2299 hours to go.
Some of this will be because one is on the processor that I use when using the computer, and some because BOINC won\'t get it right until a week or so before they finish.

The best way to check on the trickle times, is to select \"Network activity suspended\", and write down the creation time of the trickle, let it upload, then suspend access again until the next trickle, and write down that time.
You\'ll have to do this for both models.

ID: 26532 · Report as offensive     Reply Quote
old_user283420

Send message
Joined: 18 Feb 06
Posts: 6
Credit: 12,872
RAC: 0
Message 26533 - Posted: 3 Feb 2007, 2:23:10 UTC

Thanks Les, that makes sense.
As it was my 1st CPDN model I wondered why it was producing the same completion time as my Xeon but 2.5 times slower.
Many thanks
ID: 26533 · Report as offensive     Reply Quote

Questions and Answers : Windows : ClimateChange model same as BBC?

©2024 cpdn.org