I've encountered an issue, and I'd appreciate some suggestions on how to proceed. With the recent installment of Derecho on the NCAR HPC system, I needed to recompile WPS and WRF. I'm utilizing WRFV4.1.1 and WPSV4.1. After a lot of troubleshooting, I came to the following procedure for the compilation:
Setting up the environment:
module purge
module load intel-classic/2023.0.0
module load ncarcompilers/1.0.0
module load cray-mpich/8.1.25
module load craype/2.7.20
module load netcdf-mpi/4.9.2
export JASPERLIB=/glade/u/home/wrfhelp/UNGRIB_LIBRARIES/lib
export JASPERINC=/glade/u/home/wrfhelp/UNGRIB_LIBRARIES/include
Compiling WRF:
./configure --> choose option 50
./compile em_real >& log.compile --> check log.compile for sources of error
Compiling WPS:
export WRF_DIR=../WRFVX.X.X
./configure --> choose option 37
./compile >& log.compile
I am able to build the executables for WPS and WRF without issues. I am able to run ./real.exe and generate the input and boundary conditions for WRF, but upon running ./wrf.exe, I receive the following error:
FATAL CALLED FROM FILE: <stdin> LINE: 517
start_em: p_top from the namelist does not match p_top from the input file.
Upon running the command ncdump -v P_TOP wrfinput_d01 | tail -10, I receive:
ncdump: P_TOP: No such variable
It's worth noting that I'm performing a pseudo global warming experiment, so the forcing data (ERA5 on model levels) is modified prior to WPS/WRF processing. However, the original structure of the data is left undisturbed, only values for certain variables are changed. I had no issues with using the data in WPS/WRF on Cheyenne, so I'm inclined to believe the data is not the source of the problem.
I'll attach additional information to help in troubleshooting. I'm not opposed to recompiling WPS/WRF if that's the simplest solution (which it very well may be). However, it seems there may not be a well-established, reliable procedure on Derecho yet. I would appreciate some insight on this from someone who has successfully compiled and ran WPS/WRF on Derecho.
Thanks in advance!
Setting up the environment:
module purge
module load intel-classic/2023.0.0
module load ncarcompilers/1.0.0
module load cray-mpich/8.1.25
module load craype/2.7.20
module load netcdf-mpi/4.9.2
export JASPERLIB=/glade/u/home/wrfhelp/UNGRIB_LIBRARIES/lib
export JASPERINC=/glade/u/home/wrfhelp/UNGRIB_LIBRARIES/include
Compiling WRF:
./configure --> choose option 50
./compile em_real >& log.compile --> check log.compile for sources of error
Compiling WPS:
export WRF_DIR=../WRFVX.X.X
./configure --> choose option 37
./compile >& log.compile
I am able to build the executables for WPS and WRF without issues. I am able to run ./real.exe and generate the input and boundary conditions for WRF, but upon running ./wrf.exe, I receive the following error:
FATAL CALLED FROM FILE: <stdin> LINE: 517
start_em: p_top from the namelist does not match p_top from the input file.
Upon running the command ncdump -v P_TOP wrfinput_d01 | tail -10, I receive:
ncdump: P_TOP: No such variable
It's worth noting that I'm performing a pseudo global warming experiment, so the forcing data (ERA5 on model levels) is modified prior to WPS/WRF processing. However, the original structure of the data is left undisturbed, only values for certain variables are changed. I had no issues with using the data in WPS/WRF on Cheyenne, so I'm inclined to believe the data is not the source of the problem.
I'll attach additional information to help in troubleshooting. I'm not opposed to recompiling WPS/WRF if that's the simplest solution (which it very well may be). However, it seems there may not be a well-established, reliable procedure on Derecho yet. I would appreciate some insight on this from someone who has successfully compiled and ran WPS/WRF on Derecho.
Thanks in advance!