After a long struggle to figure out why I couldn't get MPAS-A to run with the default PNetcdf I/O but successfully with Netcdf-3 serial I/O, I have determined that the ultimate issue is that the conda version of parallelio I am using was built without PNetcdf support, as below. 2.5.10 is similar. I cannot figure out who is uploading this package; it's coming out of a general Conda pipeline account, but I would humbly request that they would build ParallelIO with such functionality, so that it doesn't break the default input/output mechanism in MPAS.
PIO Configuration Summary
==============================
General
PIO Version: 2.6.0
Configured On: Fri May 19 13:18:37 UTC 2023
Host System: x86_64-Linux-5.15.0-1037-azure
Build Directory: /home/conda/feedstock_root/build_artifacts/parallelio_1684502184942/work/build
Install Prefix: /home/nheavens/.conda/envs/mpas_test3
PnetCDF Support: no
SZIP Write Support: no
Parallel Filters: yes
NetCDF/HDF5 Support: yes
NetCDF/HDF5 Par I/O: yes
NetCDF Integration: no
PIO Logging: no
MPIEXEC: mpiexec
Fortran: yes
After a long struggle to figure out why I couldn't get MPAS-A to run with the default PNetcdf I/O but successfully with Netcdf-3 serial I/O, I have determined that the ultimate issue is that the conda version of parallelio I am using was built without PNetcdf support, as below. 2.5.10 is similar. I cannot figure out who is uploading this package; it's coming out of a general Conda pipeline account, but I would humbly request that they would build ParallelIO with such functionality, so that it doesn't break the default input/output mechanism in MPAS.
General
PIO Version: 2.6.0 Configured On: Fri May 19 13:18:37 UTC 2023 Host System: x86_64-Linux-5.15.0-1037-azure Build Directory: /home/conda/feedstock_root/build_artifacts/parallelio_1684502184942/work/build Install Prefix: /home/nheavens/.conda/envs/mpas_test3
Compiling Options
C Compiler: GNU 12.2.0 CFLAGS: -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -ffunction-sections -pipe -isystem /home/nheavens/.conda/envs/mpas_test3/include -fdebug-prefix-map=/home/conda/feedstock_root/build_artifacts/parallelio_1 684502184942/work=/usr/local/src/conda/parallelio-2.6.0 -fdebug-prefix-map=/home/nheavens/.conda/envs/mpas_test3=/usr/local/src/conda-prefix -std=gnu99
CPPFLAGS: -DLINUX -DCPRGNU -DLOGGING -I/home/conda/feedstock_root/build_artifacts/parallelio_1684502184942/work/build -I/home/conda/feedstock_root/build_artifacts/parallelio_1684502184942/work/src/clib -I/home/conda/feedstock_root/build_artifacts/paralle lio_1684502184942/work/build/src/clib -I/home/conda/feedstock_root/build_artifacts/parallelio_1684502184942/work/src/clib -I/home/conda/feedstock_root/build_artifacts/parallelio_1684502184942/work/src/clib/../ncint -I/home/nheavens/.conda/envs/mpas_test3/include LDFLAGS: -Wl,-O2 -Wl,--sort-common -Wl,--as-needed -Wl,-z,relro -Wl,-z,now -Wl,--disable-new-dtags -Wl,--gc-sections -Wl,--allow-shlib-undefined -Wl,-rpath,/home/nheavens/.conda/envs/mpas_test3/lib -Wl,-rpath-link,/home/nheavens/.conda/envs/mpas_test3/li b -L/home/nheavens/.conda/envs/mpas_test3/lib Shared Library: yes Static Library: no Extra libraries:
Fortran Compiler: GNU 12.2.0 FFLAGS: -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -ffunction-sections -pipe -isystem /home/nheavens/.conda/envs/mpas_test3/include -fdebug-prefix-map=/home/conda/feedstock_root/build_artifacts/parallelio_1 684502184942/work=/usr/local/src/conda/parallelio-2.6.0 -fdebug-prefix-map=/home/nheavens/.conda/envs/mpas_test3=/usr/local/src/conda-prefix -fallow-argument-mismatch FCFLAGS:
More Fortran Flags: -DLINUX -DCPRGNU -DNO_C_SIZEOF -I/home/conda/feedstock_root/build_artifacts/parallelio_1684502184942/work/build -I/home/conda/feedstock_root/build_artifacts/parallelio_1684502184942/work/src/flib -I/home/conda/feedstock_root/build_artifacts/par allelio_1684502184942/work/build/src/flib -I/home/nheavens/.conda/envs/mpas_test3/include
Features
PnetCDF Support: no SZIP Write Support: no Parallel Filters: yes NetCDF/HDF5 Support: yes NetCDF/HDF5 Par I/O: yes NetCDF Integration: no PIO Logging: no MPIEXEC: mpiexec Fortran: yes