I am obtaining error running MCIP v5.1 on benchmark wrfv3.8 input files using the gcc compiler.
MCIP V5.1 FROZEN 11/21/2019
==============================================================================
- SUBROUTINE SETUP_WRFEM - READING WRF HEADER
WRF GRID DIMENSIONS (X,Y,Z) 103 83 35
**********************************************************************
*** SUBROUTINE: SETUP_WRFEM
*** ERROR INQUIRING ABOUT VAR IN WRF FILE
*** VARIABLE = Times
*** NCF: NetCDF: Invalid argument
**********************************************************************
*** ERROR ABORT in subroutine SETUP_WRFEM
ABNORMAL TERMINATION IN SETUP_WRFEM
Date and time 0:00:00 July 2, 2011 (2011183:000000)
Error running mcip.
using gcc
gcc --version
gcc (GCC) 8.3.1 20191121 (Red Hat 8.3.1-5)
netcdf-c-4.7.3
netcdf-fortran-4.4.2
Using the following source for ioapi
I can run the same code successfully using the intel compiler.
Liz
ncdump -h subset_wrfout_d01_2016-06-30_00:00:0 > subset_wrfout_d01_2016-06-30_00:00:0.txt
Top part of the ncdump -h output file shows that Times is a character variable.
netcdf subset_wrfout_d01_2016-06-30_00\:00\:00 {
dimensions:
Time = UNLIMITED ; // (24 currently)
DateStrLen = 19 ;
south_north = 82 ;
west_east = 102 ;
bottom_top = 35 ;
bottom_top_stag = 36 ;
soil_layers_stag = 2 ;
west_east_stag = 103 ;
south_north_stag = 83 ;
land_cat_stag = 40 ;
variables:
char Times(Time, DateStrLen) ;
float XLAT(Time, south_north, west_east) ;
XLAT:FieldType = 104 ;
XLAT:MemoryOrder = "XY " ;
XLAT:description = "LATITUDE, SOUTH IS NEGATIVE" ;
XLAT:units = "degree_north" ;
XLAT:stagger = "" ;
XLAT:coordinates = "XLONG XLAT" ;
Times has always been a character variable. Nothing has changed about it.
I suspect something having to do with gcc. Was netCDF (the library and the module files) compiled using the same version of gcc used to compile MCIP?
Hi Chris,
Yes, I used the same compiler for all of the libraries and MCIP.
Liz
@lizadams –
Two questions:
- Did you also encounter the same error with MCIPv5.0 with the same versions of gcc and netCDF?
- Did you upgrade netCDF since you tested MCIPv5.0?
There could be some sensitivities that were introduced in netCDF which may have been corrected with a subsequent version of netCDF.
See https://github.com/Unidata/netcdf-c/issues/657
Happy to help dig further, if necessary.
–Tanya