[cam-users] CGD Forums

[phpbb at cgd.ucar.edu]

Dear cam-users, 

As you requested, here is the latest digest of messages posted on CGD Forums forums. Please come and join the discussion!


<<  Why do different results come from the same CCM3 code with d, http://bb.cgd.ucar.edu//viewtopic.php?t=23 >>

PosterpjrPostedSat Oct 09, 2004 4:12 pm MDT, http://bb.cgd.ucar.edu//viewtopic.php?p=62#62
Message: Why do different results come from the same CCM3 code with different platform? 
<br />
<br />   
<br />  I've found a strange thing recently, i.e., the model
<br />results are quite different with the same both CCM3
<br />code and data set running in different SGI platform,
<br />one is on origin300, and the other is one sgi3800. 
<br /> I wonder where these conflications come
<br />from. Can anybody help me? Thanks in advance.
<br />
<br />You need to tell us a little more. Do you mean that the climate is different, or that two solutions are diverging after a few days when run on the two platforms?  
<br />
<br />At some point or another we will need to make an FAQ entry on criteria useful in determining whether two long climate simulations are different. For the moment, I will just say that it is almost impossible to state definitively that two simulations are different if you are comparing averages of runs of less than 5 years duration.
<br />
<br />If your quest!
ion is about simulations which differ after a few days then the answer is that it is quite likely that:
<br />
<br />1) you are running with two different versions of the compiler 
<br />
<br />2) the compiler flags (which are frequently hidden from the user) are different for the two platforms. 
<br />
<br />The result of these items are that the model simulations will  differ at roundoff after the first time step. From then on you  will see the simulation differences grow in time according to  the classic solution differences seen in weather prediction and chaos theory.
<br />
<br />Phil

------------------------------
PosterAnonymousPostedSun Oct 10, 2004 3:26 am MDT, http://bb.cgd.ucar.edu//viewtopic.php?p=64#64
Message: Phil,
<br /> The details about my problem are the followings,
<br /> I copy the same build.test.csh, Makefile and the CCM3.6.6 codes to sgi3800 and origin300, and with the same initial data and SST data, I've run the model in the two machines for 6 days, and when I display the difference of  the two outcomes by GrADS, I find the differece amplitude increases rapidly from the order of 0.001 to -0.1, taking t(temperature) for example.
<br /> According to your advices, the NetCDF libs are same, and since the Makefiles are also same, the compile opinions should be same, too. So now I'm puzzled about this problem, and wondering which model results I should believe.
<br /> Hope your suggestions could help a lot.
<br />
<br />SUN

------------------------------




Format TEXT
Show Message Text YES
Show My Messages YES
Digest Frequency DAY
Show only new messages since last time I logged in FALSE
Send empty digests NO
Time of day to send digest 9 AM
Maximum characters per message in digest 32000