CI updates (#404)

* updated scripts for using amazon cluster

* cleaning up parallel works scripts

Author: vithikashah001

.github/.parallelworks/checkout.sh

@@ -1,4 +1,4 @@
-#!/bin/sh -xe
+#!/bin/sh -e
 
 ##############################################################################
 ## User set up variables
@@ -55,5 +55,6 @@ mkdir -p ${logDir}
 ## clone code
 cd ${testDir}
 git clone --recursive https://github.com/NOAA-GFDL/SHiELD_build.git && cd SHiELD_build && ./CHECKOUT_code |& tee ${logDir}/checkout.log
+
 ## Check out the PR
 cd ${testDir}/SHiELD_SRC/GFDL_atmos_cubed_sphere && git fetch origin ${branch}:toMerge && git merge toMerge

.github/.parallelworks/compile.sh

@@ -1,4 +1,4 @@
-#!/bin/sh -xe
+#!/bin/sh -e
 
 ##############################################################################
 ## User set up variables

.github/.parallelworks/run_test.sh

@@ -1,4 +1,4 @@
-#!/bin/bash -xe
+#!/bin/bash -e
 ulimit -s unlimited
 ##############################################################################
 ## User set up variables
@@ -9,7 +9,7 @@ intelVersion=2023.2.0
 ##############################################################################
 ## HPC-ME container
 container=/contrib/containers/noaa-intel-prototype_2023.09.25.sif
-container_env_script=/contrib/containers/load_spack_noaa-intel-mlong.sh
+container_env_script=/contrib/containers/load_spack_noaa-intel.sh
 
 #Parse Arguments
 branch=main
@@ -50,10 +50,16 @@ echo "test is $testname"
 
 ## Set up the directories
 MODULESHOME=/usr/share/lmod/lmod
+source $MODULESHOME/init/sh
+#export MODULEPATH=/mnt/shared/manual_modules:/usr/share/modulefiles/Linux:/usr/share/modulefiles/Core:/usr/share/lmod/lmod/modulefiles/Core:/apps/modules/modulefiles:/apps/modules/modulefamilies/intel
+#module load intel/2022.1.2
+#module load impi/2022.1.2
+module use -a /usr/share/Modules/modulefiles /opt/intel/impi/2019.5.281/intel64/modulefiles /apps/modules/modulefiles
+module load intelmpi
+module load nccmp
 testDir=${dirRoot}/${intelVersion}/GFDL_atmos_cubed_sphere/${branch}/${commit}
 logDir=${testDir}/log
 baselineDir=${dirRoot}/baselines/intel/${intelVersion}
-
 ## Run the CI Test
 # Define the builddir testscriptdir and rundir BUILDDIR is used by test scripts 
 # Set the BUILDDIR for the test script to use
@@ -65,15 +71,23 @@ runDir=${BUILDDIR}/CI/BATCH-CI
 cd ${testscriptDir}
 set -o pipefail
 # Execute the test piping output to log file
-./${testname} " --partition=compute --mpi=pmi2 --job-name=${commit}_${testname} singularity exec -B /contrib -B /apps ${container} ${container_env_script}" |& tee ${logDir}/run_${testname}.log
-
+./${testname} " --partition=compute --mpi=pmi2 --job-name=${commit}_${testname} singularity exec -B /contrib -B /usr/lib64/libpmi2.so ${container} ${container_env_script}" |& tee ${logDir}/run_${testname}.log
 ## Compare Restarts to Baseline
-source $MODULESHOME/init/sh
-export MODULEPATH=/mnt/shared/manual_modules:/usr/share/modulefiles/Linux:/usr/share/modulefiles/Core:/usr/share/lmod/lmod/modulefiles/Core:/apps/modules/modulefiles:/apps/modules/modulefamilies/intel
-module load intel/2022.1.2
-module load netcdf
-module load nccmp
-for resFile in `ls ${baselineDir}/${testname}`
-do
-  nccmp -d ${baselineDir}/${testname}/${resFile} ${runDir}/${testname}/RESTART/${resFile}
-done
+#The following tests are not expectred to have run-to-run reproducibility:
+#d96_2k.solo.bubble
+#d96_2k.solo.bubble.n0
+#d96_2k.solo.bubble.nhK
+if [[ ${testname} == "d96_2k.solo.bubble" || ${testname} == "d96_2k.solo.bubble.n0" || ${testname} == "d96_2k.solo.bubble.nhK" ]]
+  then
+    echo "${testname} is not expected to reproduce so answers were not compared"
+  else
+    #source $MODULESHOME/init/sh
+    #export MODULEPATH=/mnt/shared/manual_modules:/usr/share/modulefiles/Linux:/usr/share/modulefiles/Core:/usr/share/lmod/lmod/modulefiles/Core:/apps/modules/modulefiles:/apps/modules/modulefamilies/intel
+    #module load intel/2022.1.2
+    #module load netcdf
+    for resFile in `ls ${baselineDir}/${testname}`
+    do
+      echo "comparing ${runDir}/${testname}/RESTART/${resFile}"
+      nccmp -d ${baselineDir}/${testname}/${resFile} ${runDir}/${testname}/RESTART/${resFile}
+    done
+fi