[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!


<<  Port Validation in HP (formerly Compaq) Tru64, http://bb.cgd.ucar.edu//viewtopic.php?t=77 >>

PostereatonPostedTue Dec 14, 2004 3:49 pm MST, http://bb.cgd.ucar.edu//viewtopic.php?p=246#246
Message: &gt; I agree with you. The port should be validated using a full climate
<br />&gt; simulation using WACCM. Nonetheless, I think, some criteria to ensure
<br />&gt; the validity of the port is still needed in case of using the full climate
<br />&gt; simulation.
<br />
<br />I believe the criteria to validate that two 20 year simulations are samples
<br />from the same climate is an open research question (and possibly not a well
<br />posed one).  In practice one focuses on comparing those aspects of the
<br />climate simulations that are of particular interest.
<br />
<br />The value of the perturbation growth test is that it provides a simple way
<br />to compare two simulations that are expected to differ only due to the
<br />growth of machine roundoff size differences.  This validation is extremely
<br />useful for catching obvious problems with the port, before beginning the
<br />much more difficult and expensive process of running long simulations and
<br /!
 >doing a full climate validation using appropriate scientific criteria.

------------------------------
PostersunnyabcPostedWed Dec 15, 2004 1:31 am MST, http://bb.cgd.ucar.edu//viewtopic.php?p=250#250
Message: Thank you for reply
<br />
<br />Sorry, I have another question. This will be the last one.
<br />
<br />As far as I know, NCAR CGD provides officially CAM ported in the three 
<br />system (IBM, SGI, and LINUX). 
<br />
<br />What is the CGD's criteria to confirm whether the model is running 
<br />correctly in those systems? Are time-averaged fields and the perturbation
<br />growth tests used? Is the transience of variables also considered?
<br />
<br />And, is the port validated by comparing results with those obtained from
<br />IBM system?

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

<<  performance on earth simulator, http://bb.cgd.ucar.edu//viewtopic.php?t=78 >>

PosterghanPostedTue Dec 14, 2004 5:03 pm MST, http://bb.cgd.ucar.edu//viewtopic.php?p=247#247
Message: In the course of preparing a proposal for a new machine here at PNNL, I noticed that the AGCM for Earth Simulator (AFES) achieves a performance of 65% of theoretical peak performance of the Earth Simulator. That is much higher than the performance we typically see for CAM on IBM and other scalar systems, which is typically 5-10% of peak. Itis not clear whether this higher performance is due to the ES hardware or the AFES software. Does the AFES minimize use of cache so that memory bandwidth does not limit performance like it does for CAM on scalar systems? Or is memory bandwidth less of an issue for the ES? 
<br />
<br />We can answer this question by looking at the performance of CAM or CCSM on the ES. So my specific question is what is the performance of CAM or CCSM, as % of theoretical peak, on the ES? If it is high then we can conclude that the high performance of AFES is due to the ES hardware. But if it is low then perhaps CAM could be modified to perform more!
  efficiently.

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




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