Main menu

Navigation

CLM error when porting from Yellowstone to Cheyenne

1 post / 0 new
mark@...
CLM error when porting from Yellowstone to Cheyenne

Now that I have finally access to Cheyenne I wanted to test some codes on it that have successfully run on Yellowstone.  The first is using cam5_4_108 and running a standard CAM5 case (FAMIPC5) at f19_f19 resolution.  The model compiles successfully on Cheyenne but crashes after the 2nd time step giving this error:

 

6: SNICAR ERROR: snow grain radius of            0  out of bounds.

6: NSTEP=            2

6: flg_snw_ice=            1

6: column:         5917  level:            0  snl(c)=           -1

6: lat=    42.6315789473683       lon=    97.5000000000000

6: h2osno(c)=   0.778360533645100

6: calling getglobalwrite with decomp_index=         5917  and clmlevel= column

6: local  column   index =         5917

5: local  column   index =         4969

6: global column   index =        37862

5: global column   index =        37846

6: global landunit index =        12326

6: global gridcell index =         3736

6: gridcell longitude    =    97.5000000000000

6: gridcell latitude     =    42.6315789473683

6: column   type         =            1

6: landunit type         =            1

6: ENDRUN:

6: ERROR in SnowSnicarMod.F90 at line 426

 

When I go through the lnd log files they are identical up to this point:

 

** cheyenne **

 

 Attempting to read global land mask from

 /glade/p/cesmdata/cseg/inputdata/share/domains/domain.lnd.fv1.9x2.5_gx1v6.09020

 6.nc

 (GETFIL): attempting to find local file domain.lnd.fv1.9x2.5_gx1v6.090206.nc

 (GETFIL): using

 /glade/p/cesmdata/cseg/inputdata/share/domains/domain.lnd.fv1.9x2.5_gx1v6.09020

 6.nc

 lat/lon grid flag (isgrid2d) is  T

 ncd_inqvid: variable LANDMASK is not on dataset

  decomp precompute numg,nclumps,seglen1,avg_seglen,nsegspc=        5663

          72 F   2.247222       35.00000

  Surface Grid Characteristics

    longitude points               =          144

    latitude points                =           96

    total number of land gridcells =         5663

  Decomposition Characteristics

    clumps per process             =            1

  gsMap Characteristics

   lnd gsmap glo num of segs      =         2708

 

** yellowstone **

 

 Attempting to read global land mask from

 /glade/p/cesmdata/cseg/inputdata/share/domains/domain.lnd.fv1.9x2.5_gx1v6.09020

 6.nc

 (GETFIL): attempting to find local file domain.lnd.fv1.9x2.5_gx1v6.090206.nc

 (GETFIL): using

 /glade/p/cesmdata/cseg/inputdata/share/domains/domain.lnd.fv1.9x2.5_gx1v6.09020

 6.nc

 lat/lon grid flag (isgrid2d) is  T

 ncd_inqvid: variable LANDMASK is not on dataset

  decomp precompute numg,nclumps,seglen1,avg_seglen,nsegspc=        5663

         384 T   1.000000       14.74740

  Surface Grid Characteristics

    longitude points               =          144

    latitude points                =           96

    total number of land gridcells =         5663

  Decomposition Characteristics

    clumps per process             =            2

  gsMap Characteristics

   lnd gsmap glo num of segs      =         5663

 

Why would cause some of those decomp precompute variables like nclumps, etc. to be different?  

 

Thanks!

Mark Branson

 

Who's new

  • Nicholas.Davis@...
  • numarsanifa@...
  • bingdian_46@...
  • mxy2832029@...
  • nthg2000@...