Support global-workflow using Rocky 8 on CSPs

env/AWSPW.env

@@ -33,7 +33,151 @@ else
     exit 2
 fi
 
-if [[ "${step}" = "fcst" ]] || [[ "${step}" = "efcs" ]]; then
+if [[ "${step}" = "prep" ]] || [[ "${step}" = "prepbufr" ]]; then
+
+    export POE="NO"
+    export BACK="NO"
+    export sys_tp="HERA"
+    export launcher_PREP="srun"
+
+elif [[ "${step}" = "prepsnowobs" ]]; then
+
+    export APRUN_CALCFIMS="${APRUN_default}"
+
+elif [[ "${step}" = "prep_emissions" ]]; then
+
+    export APRUN="${APRUN_default}"
+
+elif [[ "${step}" = "waveinit" ]] || [[ "${step}" = "waveprep" ]] || [[ "${step}" = "wavepostsbs" ]] || [[ "${step}" = "wavepostbndpnt" ]] || [[ "${step}" = "wavepostbndpntbll" ]] || [[ "${step}" = "wavepostpnt" ]]; then
+
+    export CFP_MP="YES"
+    if [[ "${step}" = "waveprep" ]]; then export MP_PULSE=0 ; fi
+    export wavempexec=${launcher}
+    export wave_mpmd=${mpmd_opt}
+
+elif [[ "${step}" = "atmanlvar" ]]; then
+
+    export NTHREADS_ATMANLVAR=${NTHREADSmax}
+    export APRUN_ATMANLVAR="${APRUN_default} --cpus-per-task=${NTHREADS_ATMANLVAR}"
+
+elif [[ "${step}" = "atmensanlletkf" ]]; then
+
+    export NTHREADS_ATMENSANLLETKF=${NTHREADSmax}
+    export APRUN_ATMENSANLLETKF="${APRUN_default} --cpus-per-task=${NTHREADS_ATMENSANLLETKF}"
+
+elif [[ "${step}" = "atmensanlfv3inc" ]]; then
+
+    export NTHREADS_ATMENSANLFV3INC=${NTHREADSmax}
+    export APRUN_ATMENSANLFV3INC="${APRUN_default} --cpus-per-task=${NTHREADS_ATMENSANLFV3INC}"    
+
+elif [[ "${step}" = "aeroanlrun" ]]; then
+
+    export APRUNCFP="${launcher} -n \$ncmd ${mpmd_opt}"
+
+    export NTHREADS_AEROANL=${NTHREADSmax}
+    export APRUN_AEROANL="${APRUN_default} --cpus-per-task=${NTHREADS_AEROANL}"
+
+elif [[ "${step}" = "atmanlfv3inc" ]]; then
+
+    export NTHREADS_ATMANLFV3INC=${NTHREADSmax}
+    export APRUN_ATMANLFV3INC="${APRUN_default} --cpus-per-task=${NTHREADS_ATMANLFV3INC}"
+
+elif [[ "${step}" = "prepobsaero" ]]; then
+
+    export NTHREADS_PREPOBSAERO=${NTHREADS1}
+    export APRUN_PREPOBSAERO="${APRUN_default} --cpus-per-task=${NTHREADS_PREPOBSAERO}"
+
+elif [[ "${step}" = "snowanl" ]]; then
+
+    export NTHREADS_SNOWANL=${NTHREADSmax}
+    export APRUN_SNOWANL="${APRUN_default} --cpus-per-task=${NTHREADS_SNOWANL}"
+
+    export APRUN_APPLY_INCR="${launcher} -n 6"
+
+elif [[ "${step}" = "marinebmat" ]]; then
+
+    export APRUNCFP="${launcher} -n \$ncmd --multi-prog"
+    export APRUN_MARINEBMAT="${APRUN_default}"
+
+elif [[ "${step}" = "marinebmat" ]]; then
+
+    export APRUNCFP="${launcher} -n \$ncmd --multi-prog"
+    export APRUN_MARINEBMAT="${APRUN_default}"
+
+elif [[ "${step}" = "ocnanalrun" ]]; then
+
+    export APRUNCFP="${launcher} -n \$ncmd --multi-prog"
+
+    export APRUN_OCNANAL="${APRUN_default}"
+
+elif [[ "${step}" = "ocnanalchkpt" ]]; then
+
+    export APRUNCFP="${launcher} -n \$ncmd --multi-prog"
+
+    export APRUN_OCNANAL="${APRUN_default}"
+
+elif [[ "${step}" = "ocnanalecen" ]]; then
+
+    export NTHREADS_OCNANALECEN=${NTHREADSmax}
+    export APRUN_OCNANALECEN="${APRUN_default} --cpus-per-task=${NTHREADS_OCNANALECEN}"
+
+elif [[ "${step}" = "marineanalletkf" ]]; then
+
+    export NTHREADS_MARINEANALLETKF=${NTHREADSmax}
+    export APRUN_MARINEANALLETKF="${APRUN_default} --cpus-per-task=${NTHREADS_MARINEANALLETKF}"
+
+elif [[ "${step}" = "anal" ]] || [[ "${step}" = "analcalc" ]]; then
+
+    export MKL_NUM_THREADS=4
+    export MKL_CBWR=AUTO
+
+    export CFP_MP=${CFP_MP:-"YES"}
+    export USE_CFP=${USE_CFP:-"YES"}
+    export APRUNCFP="${launcher} -n \$ncmd ${mpmd_opt}"
+
+    export NTHREADS_GSI=${NTHREADSmax}
+    export APRUN_GSI="${APRUN_default} --cpus-per-task=${NTHREADS_GSI}"
+
+    export NTHREADS_CALCINC=${threads_per_task_calcinc:-1}
+    [[ ${NTHREADS_CALCINC} -gt ${max_threads_per_task} ]] && export NTHREADS_CALCINC=${max_threads_per_task}
+    export APRUN_CALCINC="${launcher} \$ncmd --cpus-per-task=${NTHREADS_CALCINC}"
+
+    export NTHREADS_CYCLE=${threads_per_task_cycle:-12}
+    [[ ${NTHREADS_CYCLE} -gt ${max_tasks_per_node} ]] && export NTHREADS_CYCLE=${max_tasks_per_node}
+    ntasks_cycle=${ntiles:-6}
+    export APRUN_CYCLE="${launcher} -n ${ntasks_cycle} --cpus-per-task=${NTHREADS_CYCLE}"
+
+    export NTHREADS_GAUSFCANL=1
+    ntasks_gausfcanl=${ntasks_gausfcanl:-1}
+    export APRUN_GAUSFCANL="${launcher} -n ${ntasks_gausfcanl} --cpus-per-task=${NTHREADS_GAUSFCANL}"
+
+elif [[ "${step}" = "sfcanl" ]]; then
+
+    export NTHREADS_CYCLE=${threads_per_task:-14}
+    export APRUN_CYCLE="${APRUN_default} --cpus-per-task=${NTHREADS_CYCLE}"
+
+elif [[ "${step}" = "eobs" ]]; then
+
+    export MKL_NUM_THREADS=4
+    export MKL_CBWR=AUTO
+
+    export NTHREADS_GSI=${NTHREADSmax}
+    export APRUN_GSI="${APRUN_default} --cpus-per-task=${NTHREADS_GSI}"
+
+    export CFP_MP=${CFP_MP:-"YES"}
+    export USE_CFP=${USE_CFP:-"YES"}
+    export APRUNCFP="${launcher} -n \$ncmd ${mpmd_opt}"
+
+elif [[ "${step}" = "eupd" ]]; then
+
+    export NTHREADS_ENKF=${NTHREADSmax}
+    export APRUN_ENKF="${launcher} -n ${ntasks_enkf:-${ntasks}} --cpus-per-task=${NTHREADS_ENKF}"
+
+    export CFP_MP=${CFP_MP:-"YES"}
+    export USE_CFP=${USE_CFP:-"YES"}
+    export APRUNCFP="${launcher} -n \$ncmd ${mpmd_opt}"
+
+elif [[ "${step}" = "fcst" ]] || [[ "${step}" = "efcs" ]]; then
 
     export launcher="srun --mpi=pmi2 -l"
 
@@ -52,22 +196,18 @@ elif [[ "${step}" = "waveinit" ]] || [[ "${step}" = "waveprep" ]] || [[ "${step}
 
 elif [[ "${step}" = "post" ]]; then
 
-    export NTHREADS_NP=${NTHREADS1}
-    export APRUN_NP="${APRUN_default}"
-
-    export NTHREADS_DWN=${threads_per_task_dwn:-1}
-    [[ ${NTHREADS_DWN} -gt ${max_threads_per_task} ]] && export NTHREADS_DWN=${max_threads_per_task}
-    export APRUN_DWN="${launcher} -n ${ntasks_dwn}"
-
-elif [[ "${step}" = "atmos_products" ]]; then
-
-    export USE_CFP="YES"  # Use MPMD for downstream product generation on Hera
+    export NTHREADS_UPP=${NTHREADS1}
+    export APRUN_UPP="${APRUN_default} --cpus-per-task=${NTHREADS_UPP}"
 
 elif [[ "${step}" = "oceanice_products" ]]; then
 
     export NTHREADS_OCNICEPOST=${NTHREADS1}
     export APRUN_OCNICEPOST="${launcher} -n 1 --cpus-per-task=${NTHREADS_OCNICEPOST}"
 
+elif [[ "${step}" = "atmos_products" ]]; then
+
+    export USE_CFP="YES"  # Use MPMD for downstream product generation on Hera
+
 elif [[ "${step}" = "ecen" ]]; then
 
     export NTHREADS_ECEN=${NTHREADSmax}

env/AZUREPW.env

@@ -15,6 +15,7 @@ export mpmd_opt="--multi-prog --output=mpmd.%j.%t.out"
 # Configure MPI environment
 export OMP_STACKSIZE=2048000
 export NTHSTACK=1024000000
+export UCX_TLS=ud,sm,self
 
 ulimit -s unlimited
 ulimit -a
@@ -50,6 +51,10 @@ elif [[ "${step}" = "waveinit" ]] || [[ "${step}" = "waveprep" ]] || [[ "${step}
     export wavempexec=${launcher}
     export wave_mpmd=${mpmd_opt}
 
+elif [[ "${step}" = "prep_emissions" ]]; then
+
+    export APRUN="${APRUN_default}"
+
 elif [[ "${step}" = "post" ]]; then
 
     export NTHREADS_NP=${NTHREADS1}
@@ -71,33 +76,33 @@ elif [[ "${step}" = "oceanice_products" ]]; then
 elif [[ "${step}" = "ecen" ]]; then
 
     export NTHREADS_ECEN=${NTHREADSmax}
-    export APRUN_ECEN="${APRUN}"
+    export APRUN_ECEN="${APRUN_default}"
 
     export NTHREADS_CHGRES=${threads_per_task_chgres:-12}
     [[ ${NTHREADS_CHGRES} -gt ${max_tasks_per_node} ]] && export NTHREADS_CHGRES=${max_tasks_per_node}
     export APRUN_CHGRES="time"
 
     export NTHREADS_CALCINC=${threads_per_task_calcinc:-1}
     [[ ${NTHREADS_CALCINC} -gt ${max_threads_per_task} ]] && export NTHREADS_CALCINC=${max_threads_per_task}
-    export APRUN_CALCINC="${APRUN}"
+    export APRUN_CALCINC="${APRUN_default}"
 
 elif [[ "${step}" = "esfc" ]]; then
 
     export NTHREADS_ESFC=${NTHREADSmax}
-    export APRUN_ESFC="${APRUN}"
+    export APRUN_ESFC="${APRUN_default}"
 
     export NTHREADS_CYCLE=${threads_per_task_cycle:-14}
     [[ ${NTHREADS_CYCLE} -gt ${max_tasks_per_node} ]] && export NTHREADS_CYCLE=${max_tasks_per_node}
-    export APRUN_CYCLE="${APRUN}"
+    export APRUN_CYCLE="${APRUN_default}"
 
 elif [[ "${step}" = "epos" ]]; then
 
     export NTHREADS_EPOS=${NTHREADSmax}
-    export APRUN_EPOS="${APRUN}"
+    export APRUN_EPOS="${APRUN_default}"
 
 elif [[ "${step}" = "fit2obs" ]]; then
 
     export NTHREADS_FIT2OBS=${NTHREADS1}
-    export MPIRUN="${APRUN}"
+    export MPIRUN="${APRUN_default}"
 
 fi

env/GOOGLEPW.env

@@ -45,7 +45,7 @@ if [[ "${step}" = "fcst" ]] || [[ "${step}" = "efcs" ]]; then
 
 elif [[ "${step}" = "prep_emissions" ]]; then
 
-    export APRUN
+    export APRUN="${APRUN_default}"
 
 elif [[ "${step}" = "waveinit" ]] || [[ "${step}" = "waveprep" ]] || [[ "${step}" = "wavepostsbs" ]] || [[ "${step}" = "wavepostbndpnt" ]] || [[ "${step}" = "wavepostbndpntbll" ]] || [[ "${step}" = "wavepostpnt" ]]; then
 
@@ -102,6 +102,6 @@ elif [[ "${step}" = "epos" ]]; then
 elif [[ "${step}" = "fit2obs" ]]; then
 
     export NTHREADS_FIT2OBS=${NTHREADS1}
-    export MPIRUN="${APRUN}"
+    export MPIRUN="${APRUN_default}"
 
 fi

modulefiles/module_base.noaacloud.lua

@@ -5,8 +5,11 @@ Load environment to run GFS on noaacloud
 local spack_mod_path=(os.getenv("spack_mod_path") or "None")
 prepend_path("MODULEPATH", spack_mod_path)
 
+load("gnu")
 load(pathJoin("stack-intel", (os.getenv("stack_intel_ver") or "None")))
 load(pathJoin("stack-intel-oneapi-mpi", (os.getenv("stack_impi_ver") or "None")))
+unload("gnu")
+
 load(pathJoin("python", (os.getenv("python_ver") or "None")))
 
 load(pathJoin("jasper", (os.getenv("jasper_ver") or "None")))

modulefiles/module_gwci.noaacloud.lua

@@ -2,10 +2,10 @@ help([[
 Load environment to run GFS workflow setup scripts on noaacloud
 ]])
 
-prepend_path("MODULEPATH", "/contrib/spack-stack/spack-stack-1.6.0/envs/unified-env/install/modulefiles/Core")
+prepend_path("MODULEPATH", "/contrib/spack-stack-rocky8/spack-stack-1.6.0/envs/ue-env/install/modulefiles/Core")
 
-load(pathJoin("stack-intel", os.getenv("2021.3.0")))
-load(pathJoin("stack-intel-oneapi-mpi", os.getenv("2021.3.0")))
+load(pathJoin("stack-intel", os.getenv("2021.10.0")))
+load(pathJoin("stack-intel-oneapi-mpi", os.getenv("2021.10.0")))
 
 load(pathJoin("netcdf-c", os.getenv("4.9.2")))
 load(pathJoin("netcdf-fortran", os.getenv("4.6.1")))

modulefiles/module_gwsetup.noaacloud.lua

@@ -4,17 +4,18 @@ Load environment to run GFS workflow setup scripts on noaacloud
 
 load(pathJoin("rocoto"))
 
-prepend_path("MODULEPATH", "/contrib/spack-stack/spack-stack-1.6.0/envs/unified-env/install/modulefiles/Core")
+prepend_path("MODULEPATH", "/contrib/spack-stack-rocky8/spack-stack-1.6.0/envs/ue-intel/install/modulefiles/Core")
 
-local stack_intel_ver=os.getenv("stack_intel_ver") or "2021.3.0"
-local python_ver=os.getenv("python_ver") or "3.10.3"
+load("gnu")
+local stack_intel_ver=os.getenv("stack_intel_ver") or "2021.10.0"
+local stack_mpi_ver=os.getenv("stack_mpi_ver") or "2021.10.0"
 
 load(pathJoin("stack-intel", stack_intel_ver))
-load(pathJoin("python", python_ver))
+load(pathJoin("stack-intel-oneapi-mpi", stack_mpi_ver))
+unload("gnu")
+
 load("py-jinja2")
 load("py-pyyaml")
 load("py-numpy")
-local git_ver=os.getenv("git_ver") or "1.8.3.1"
-load(pathJoin("git", git_ver))
 
 whatis("Description: GFS run setup environment")

parm/config/gefs/config.resources

@@ -41,15 +41,15 @@ case ${machine} in
     ;;
   "AWSPW")
     export PARTITION_BATCH="compute"
-    max_tasks_per_node=36
+    max_tasks_per_node=48
     ;;
   "AZUREPW")
     export PARTITION_BATCH="compute"
-    max_tasks_per_node=24
+    max_tasks_per_node=36
     ;;
   "GOOGLEPW")
     export PARTITION_BATCH="compute"
-    max_tasks_per_node=32
+    max_tasks_per_node=30
     ;;
   *)
     echo "FATAL ERROR: Unknown machine encountered by ${BASH_SOURCE[0]}"

parm/config/gefs/config.resources.AWSPW

@@ -9,3 +9,61 @@ unset memory
 for mem_var in $(env | grep '^memory_' | cut -d= -f1); do
   unset "${mem_var}"
 done
+
+step=$1
+
+case ${step} in
+  "fcst" | "efcs")
+    export PARTITION_BATCH="compute"
+    max_tasks_per_node=48
+    ;;
+
+  "arch")
+    export PARTITION_BATCH="process"
+    max_tasks_per_node=24
+    ;;
+
+  "prep_emissions")
+    export PARTITION_BATCH="process"
+    max_tasks_per_node=24
+    export ntasks=1
+    export threads_per_task=1
+    export tasks_per_node=$(( max_tasks_per_node / threads_per_task ))
+    ;;
+
+  "waveinit")
+    export PARTITION_BATCH="process"
+    max_tasks_per_node=24
+    export ntasks=12
+    export threads_per_task=1
+    export tasks_per_node=$(( max_tasks_per_node / threads_per_task ))
+    export NTASKS=${ntasks}
+    ;;
+
+  "wavepostpnt")
+    export PARTITION_BATCH="compute"
+    max_tasks_per_node=48
+    export ntasks=240
+    export threads_per_task=1
+    export tasks_per_node=$(( max_tasks_per_node / threads_per_task ))
+    export NTASKS=${ntasks}
+    ;;
+
+  "wavepostsbs" | "wavepostbndpnt" | "wavepostbndpntbll")
+    export PARTITION_BATCH="process"
+    max_tasks_per_node=24
+    export ntasks=24
+    export threads_per_task=1
+    export tasks_per_node=$(( max_tasks_per_node / threads_per_task ))
+    export NTASKS=${ntasks}
+    ;;
+
+  *)
+    export PARTITION_BATCH="process"
+    max_tasks_per_node=24
+    ;;
+
+esac
+
+export max_tasks_per_node
+