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

Running CESM1.2.0 on Local PC POSSIBLE???

A

abdulla_sakalli@jrc_ec_europa_eu

Member
Dear Forum Member,since I have a lot of problem with our server support, I'm thinking to run cesm1.2.0 on my local pc. It includes 24 CPU with each 6 cores and 12 threads. Linux X86_64 running as OS and I have openmpi installed. Is it possible to get cesm running on this machine? So far so gut, I got the model compiling and building on the machine. But the running crashes with an error "File drv_in not found, setting default values" and emory block size conversion in bytes is          1025.50
seq_flds_mod: read seq_cplflds_inparm namelist from: drv_in
At line 359 of file /Data3/ces1.2/models/drv/shr/seq_flds_mod.F90 (unit = 90, file = '^G^PESC^?')
Fortran runtime error: File 'drv_in' does not exist
--------------------------------------------------------------------------
mpirun has exited due to process rank 0 with PID 2424 on
node D01RI1302290 exiting without calling "finalize". This may
have caused other processes in the application to be
terminated by signals sent by mpirun (as reported here) If it is possible to run the model on this machine, how i can configure the NTASK and NTHREADS in the config_pes.xml file? I'll mainly run the model with ICN compset.  I hope to get any reply before I spend for trying too much time.
 Thanks for zour help in advance. Best wishes
 
J

jedwards

CSEG and Liaisons
Staff member
You should be able to get CESM running on your machine.   As a first pass I would try setting NTASKS_*  = 4 NTHREADS_* = 1 in env_mach_pes.xml once that is working you can experiment with other pe counts.   The drv_in file is put into the run directory by the script preview_namelists check first to see that that script is working correctly.   If the drv_in is in the right place then look into the $(CASE).run script to make sure that you are in the right directory when the model starts.  
 
S

santos

Member
If you are using PGI, it looks like this error can sometimes occur due to using a wrong "mpirun" or "mpiexec" command. You should check to make sure that the mpirun command is correct and matches your env_mach_pes settings.
 
A

abdulla_sakalli@jrc_ec_europa_eu

Member
Dear all,I already fixed the problem with drv_in. For all who have the same problem, as jadvard wrote, you have to be sure that you are in the $RUNDIR when you start to run the model with mpirun or mpiexec. Now, i get some memory problem which causeb by mpi. The reason could be that my NetCDF did not installed with the --enable-parallel and hdf5. I'll try to install the library again. Best wishes
 
A

abdulla_sakalli@jrc_ec_europa_eu

Member
for all people who want to run the cesm1.2.0 on local pc. The configuration could be done like this:in config_machines.xml 
        Linux 64bit                                
        gnu    
        mpich2               
        /Data2/EXE/$CASE/run                      
        /Data2/EXE/$CASE/bld                    
        /Data2/inputdata          
        LINUX                             
        2
         32


here, the model cannot compiled with openmpi MPILIBS, I got lot of errors. The config_compiler.xml could also be configured as: 
 
  -DFORTRANUNDERSCORE -DNO_R16
  -fopenmp
  -fopenmp
  -fopenmp
  -g -Wall
 
  -O2 -fconvert=big-endian -ffree-line-length-none -ffixed-line-length-none
  -O0
  gfortran
  gcc
  mpif90
  mpicc 
  /home/usernamelocal
  -L/home/username/local/lib -lnetcdff -L/home/sakalab/local/lib -lnetcdf -lhdf5_hl -lhdf5 -lm -lz -ldl


I also installed the netcdf and hdf5 libs with zlib. I did not configured the config_pes.xml file. I just run the model on 64 processors. the most important thing is to configure the mkbatch file for your pc. Please find the file in attachment. I have a Linux 3.8.0-32-generic #47-Ubuntu SMP Tue Oct 1 22:35:23 UTC 2013 x86_64 x86_64 x86_64 GNU/Linux with 24 CPU a 6 cores. I hope, the informations can help you. I also did a testrun for 3 month case B the result is like:create_newcase -case /Data2/Case/TEST_B -mach LocalPC -compset B -res f19_g16---------------- CCSM TIMING PROFILE ---------------------

  Case        : TEST_B
  LID         : 131025-153842
  Machine     : LocalPC
  Caseroot    : /Data2/Case/TEST_B
  Timeroot    : /Data2/Case/TEST_B/Tools
  CCSM User   : username
  CCSM Tag    : cesm1_2_0  (best guess)
  Curr Date   : Fr 25. Okt 15:38:45 CEST 2013

  grid        : a%1.9x2.5_l%1.9x2.5_oi%gx1v6_r%r05_m%gx1v6_g%null_w%null
  compset     : 2000_CAM4_CLM40%SP_CICE_POP2_RTM_SGLC_SWAV
  run_type    : startup, continue_run = FALSE (inittype = TRUE)
  stop_option : nmonths, stop_n = 3
  run_length  : 90 days (89 for ocean)



  component       comp_pes    root_pe   tasks  x threads instances (stride)
  ---------        ------     -------   ------   ------  ---------  ------ 
  cpl = cpl        64          0        64     x 1       1      (1     )
  glc = sglc       64          0        64     x 1       1      (1     )
  wav = swav       64          0        64     x 1       1      (1     )
  lnd = clm        64          0        64     x 1       1      (1     )
  rof = rtm        64          0        64     x 1       1      (1     )
  ice = cice       64          0        64     x 1       1      (1     )
  atm = cam        64          0        64     x 1       1      (1     )
  ocn = pop2       64          0        64     x 1       1      (1     )

  total pes active           : 64
  pes per node               : 32
  pe count for cost estimate : 64

 Overall Metrics:
    Model Cost:            1509.59   pe-hrs/simulated_year
    Model Throughput:         1.02   simulated_years/day

    Init Time   :      72.765 seconds
    Run Time    :   20937.872 seconds      232.643 seconds/day
    Final Time  :       3.058 seconds

    Actual Ocn Init Wait Time     :       0.001 seconds
    Estimated Ocn Init Run Time   :     105.307 seconds
    Estimated Run Time Correction :     105.306 seconds
      (This correction has been applied to the ocean and total run times)

Runs Time in total seconds, seconds/model-day, and model-years/wall-day
CPL Run Time represents time in CPL pes alone, not including time associated with data exchange with other components

    TOT Run Time:   20937.872 seconds      232.643 seconds/mday         1.02 myears/wday
    LND Run Time:     332.932 seconds        3.699 seconds/mday        63.99 myears/wday
    ROF Run Time:      45.830 seconds        0.509 seconds/mday       464.85 myears/wday
    ICE Run Time:    3206.565 seconds       35.629 seconds/mday         6.64 myears/wday
    ATM Run Time:    5643.532 seconds       62.706 seconds/mday         3.77 myears/wday
    OCN Run Time:    9477.607 seconds      105.307 seconds/mday         2.25 myears/wday
    GLC Run Time:       0.000 seconds        0.000 seconds/mday         0.00 myears/wday
    WAV Run Time:       0.000 seconds        0.000 seconds/mday         0.00 myears/wday
    CPL Run Time:     716.618 seconds        7.962 seconds/mday        29.73 myears/wday
    CPL COMM Time:   2390.852 seconds       26.565 seconds/mday         8.91 myears/wday


enjoy it
 
You must log in or register to reply here.
Top