Scheduled Downtime
On Tuesday 24 October 2023 @ 5pm MT the forums will be in read only mode in preparation for the downtime. On Wednesday 25 October 2023 @ 5am MT, this website will be down for maintenance and expected to return online later in the morning.
Normal Operations
The forums are back online with normal operations. If you notice any issues or errors related to the forums, please reach out to help@ucar.edu

Exp setup and PES layout

C

climatemodelling@gmx_com

Member
Dear Forum,

I would like to run a CAM4 standalone experiment with prescribed time-varying SST forcing.
I was thinking of using either the F_1850 or F_AMIP_CN compset as a staring point.
The advantage of the AMIP experiment would be that it is easy to just replace the time-varying SST forcing.
Is it also possible to use the F_1850 setup and prescribe a SST boundary forcing file that is not a climatology?
I am asking because preferably in my experiment all atmospheric variables should be constant at per-industrial conditions but the SST and sea-ice forcing should vary in time (e.g. a 50 year integration).

Further I want to run the 2 degree resolution grid on 1 node with 24 processors, is this outline correct?


Code: 
<entry id="NTASKS_ATM"   value="24"  />
<entry id="NTHRDS_ATM"   value="1"  />
<entry id="ROOTPE_ATM"   value="0"  />

<entry id="NTASKS_LND"   value="24"  />
<entry id="NTHRDS_LND"   value="1"  />
<entry id="ROOTPE_LND"   value="0"  />

<entry id="NTASKS_ICE"   value="1"  />
<entry id="NTHRDS_ICE"   value="1"  />
<entry id="ROOTPE_ICE"   value="0"  />

<entry id="NTASKS_OCN"   value="1"  />
<entry id="NTHRDS_OCN"   value="1"  />
<entry id="ROOTPE_OCN"   value="0"  />

<entry id="NTASKS_CPL"   value="24"  />
<entry id="NTHRDS_CPL"   value="1"  />
<entry id="ROOTPE_CPL"   value="0"  />

<entry id="NTASKS_GLC"   value="1"  />
<entry id="NTHRDS_GLC"   value="1"  />
<entry id="ROOTPE_GLC"   value="0"  />

<entry id="PSTRID_ATM"   value="1"  />
<entry id="PSTRID_LND"   value="1"  />
<entry id="PSTRID_ICE"   value="1"  />
<entry id="PSTRID_OCN"   value="1"  />
<entry id="PSTRID_CPL"   value="1"  />
<entry id="PSTRID_GLC"   value="1"  />

<entry id="TOTALPES"   value="24"  />
<entry id="PES_LEVEL"   value="1r"  />
<entry id="MAX_TASKS_PER_NODE"   value="24"  />
<entry id="PES_PER_NODE"   value="$MAX_TASKS_PER_NODE"  />
<entry id="BATCH_PES"   value="0"  />
<entry id="CCSM_PCOST"   value="0"  />
<entry id="CCSM_TCOST"   value="0"  />
<entry id="CCSM_ESTCOST"   value="0"  />

<entry id="CICE_AUTO_DECOMP"   value="true"  />


And running a 1 degree grid on e.g. 2 nodes with 24 npes should be?


Code: 
<entry id="NTASKS_ATM"   value="48"  />
<entry id="NTHRDS_ATM"   value="1"  />
<entry id="ROOTPE_ATM"   value="0"  />

<entry id="NTASKS_LND"   value="48"  />
<entry id="NTHRDS_LND"   value="1"  />
<entry id="ROOTPE_LND"   value="0"  />

<entry id="NTASKS_ICE"   value="1"  />
<entry id="NTHRDS_ICE"   value="1"  />
<entry id="ROOTPE_ICE"   value="0"  />

<entry id="NTASKS_OCN"   value="1"  />
<entry id="NTHRDS_OCN"   value="1"  />
<entry id="ROOTPE_OCN"   value="0"  />

<entry id="NTASKS_CPL"   value="48"  />
<entry id="NTHRDS_CPL"   value="1"  />
<entry id="ROOTPE_CPL"   value="0"  />

<entry id="NTASKS_GLC"   value="1"  />
<entry id="NTHRDS_GLC"   value="1"  />
<entry id="ROOTPE_GLC"   value="0"  />

<entry id="PSTRID_ATM"   value="1"  />
<entry id="PSTRID_LND"   value="1"  />
<entry id="PSTRID_ICE"   value="1"  />
<entry id="PSTRID_OCN"   value="1"  />
<entry id="PSTRID_CPL"   value="1"  />
<entry id="PSTRID_GLC"   value="1"  />

<entry id="TOTALPES"   value="48"  />
<entry id="PES_LEVEL"   value="1r"  />
<entry id="MAX_TASKS_PER_NODE"   value="24"  />
<entry id="PES_PER_NODE"   value="$MAX_TASKS_PER_NODE"  />
<entry id="BATCH_PES"   value="0"  />
<entry id="CCSM_PCOST"   value="0"  />
<entry id="CCSM_TCOST"   value="0"  />
<entry id="CCSM_ESTCOST"   value="0"  />

<entry id="CICE_AUTO_DECOMP"   value="true"  />


Thanks!!!
 
E

eaton

CSEG and Liaisons
I'm also not sure of the best path. But I'd be tempted to start from the AMIP compset and modify it to use the 1850 climatology datasets that are part of the 1850 control simulations. CAM gets its settings for the F_1850 compset from the file bld/namelist_files/use_cases/1850_cam4.xml. It gets settings for the F_AMIP_CN compset from bld/namelist_files/use_cases/1850-2005_cam4.xml. The AMIP file could be modified by replacing the information about the solar, GHG, aerosol, and ozone data by the specifications from the 1850 file.

The F compsets are typically set up with all components running on the same set of processors (so there is no concurrancy in the setup). In that case the NTASKS_* values should be the same for all the components. I think what you suggested would work, but I'd expect the performance to be less than if you give the OCN and ICE components the same number of tasks as ATM and LND. (GLC doesn't matter as it's not active in the F compsets).
 
C

climatemodelling@gmx_com

Member
Thanks Eaton! I think i will be able to set up the experiment with this information.
I assume running a the same experiment but with prescribed SST + ocean mixed layer dynamics can be achieved in a similar way but starting with an E or F compset?
 
C

climatemodelling@gmx_com

Member
eaton said:
I'm also not sure of the best path. But I'd be tempted to start from the AMIP compset and modify it to use the 1850 climatology datasets that are part of the 1850 control simulations. CAM gets its settings for the F_1850 compset from the file bld/namelist_files/use_cases/1850_cam4.xml. It gets settings for the F_AMIP_CN compset from bld/namelist_files/use_cases/1850-2005_cam4.xml. The AMIP file could be modified by replacing the information about the solar, GHG, aerosol, and ozone data by the specifications from the 1850 file.

The F compsets are typically set up with all components running on the same set of processors (so there is no concurrancy in the setup). In that case the NTASKS_* values should be the same for all the components. I think what you suggested would work, but I'd expect the performance to be less than if you give the OCN and ICE components the same number of tasks as ATM and LND. (GLC doesn't matter as it's not active in the F compsets).
Click to expand...

I am able to build the F_1850 compset without any problems.
Then I tried starting out with the F_1850 compset and only doing the following changes to switch from the climatological SST boundary forcing file to the full SST file. However, after these changes the land model build crashes. Are there any namelist options that I am missing to perform the switch?


Code: 
<!--"Run start date (yyyy-mm-dd). Only used for startup or hybrid runs (char) " -->
<entry id="RUN_STARTDATE"   value="1950-01-01"  />

<!--"Sets sst/ice_cov filename for amip runs, only used in F compset (char) " -->
<entry id="DOCN_SSTDATA_FILENAME"   value="$DIN_LOC_ROOT/atm/cam/sst/sst_HadOIBl_bc_1.9x2.5_1850_2008_c100127.nc"  />

<!--"Sets year start of sst/ice_cov for amip runs, only used in F compset (integer) " -->
<entry id="DOCN_SSTDATA_YEAR_START"   value="1850"  />

<!--"Sets year end of sst/ice_cov for amip runs, only used in F compset (integer) " -->
<entry id="DOCN_SSTDATA_YEAR_END"   value="2008"  />
 
You must log in or register to reply here.
Top