From 147da76519620a83c53b1c7d876fa54395b2bdab Mon Sep 17 00:00:00 2001
From: "g.mudalige@warwick.ac.uk" <g.mudalige@warwick.ac.uk>
Date: Mon, 12 Feb 2024 19:12:50 +0000
Subject: [PATCH] Update test scripts

---
 apps/c/CloverLeaf/test.sh                |   8 +-
 apps/c/TeaLeaf/test.sh                   |   5 +-
 apps/c/adi/test.sh                       |  29 ++-
 apps/c/lowdim_test/test.sh               | 175 +++++++++---------
 apps/c/mb_shsgc/Max_datatransfer/test.sh | 214 +++++++++++------------
 5 files changed, 221 insertions(+), 210 deletions(-)

diff --git a/apps/c/CloverLeaf/test.sh b/apps/c/CloverLeaf/test.sh
index eeeac0321..f803e5bfc 100755
--- a/apps/c/CloverLeaf/test.sh
+++ b/apps/c/CloverLeaf/test.sh
@@ -12,8 +12,6 @@ export SOURCE_AMD_HIP=source_amd_rocm-5.4.3_pythonenv
 #export TELOS=TRUE
 #export KOS=TRUE
 
-#<<comment
-
 if [[ -v TELOS || -v KOS ]]; then
 
 #============================ Test with Intel Classic Compilers==========================================
@@ -23,8 +21,8 @@ source ../../scripts/$SOURCE_INTEL
 #make -j -B
 make clean
 make
-cd $OPS_INSTALL_PATH/../apps/c/CloverLeaf
 
+cd $OPS_INSTALL_PATH/../apps/c/CloverLeaf
 make clean
 rm -f .generated
 make IEEE=1 cloverleaf_dev_seq cloverleaf_dev_mpi cloverleaf_seq cloverleaf_tiled cloverleaf_openmp cloverleaf_mpi \
@@ -118,8 +116,6 @@ echo "All Intel classic complier based applications ---- PASSED"
 
 fi
 
-#comment
-
 if [[ -v TELOS ]]; then
 
 #============================ Test with Intel SYCL Compilers==========================================
@@ -224,7 +220,7 @@ rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
 rm -f clover.out
 
 if [[ -v CUDA_INSTALL_PATH ]]; then
-make IEEE=1 cloverleaf_cuda cloverleaf_mpi_cuda cloverleaf_mpi_cuda_tiled 
+make IEEE=1 cloverleaf_cuda cloverleaf_mpi_cuda cloverleaf_mpi_cuda_tiled
 #cloverleaf_mpi_openacc_tiled cloverleaf_openacc cloverleaf_mpi_openacc \
 
 echo '============> Running CUDA'
diff --git a/apps/c/TeaLeaf/test.sh b/apps/c/TeaLeaf/test.sh
index 0188d2d04..2813e8c4c 100755
--- a/apps/c/TeaLeaf/test.sh
+++ b/apps/c/TeaLeaf/test.sh
@@ -118,6 +118,7 @@ fi
 echo "All Intel classic complier based applications ---- PASSED"
 
 
+
 if [[ -v TELOS ]]; then
 
 #============================ Test with Intel SYCL Compilers==========================================
@@ -141,7 +142,7 @@ rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
 rm perf_out
 
 echo '============> Running MPI+SYCL on CPU'
-$MPI_INSTALL_PATH/bin/mpirun -np 20 ./tealeaf_mpi_sycl OPS_CL_DEVICE=0 OPS_BLOCK_SIZE_X=256 OPS_BLOCK_SIZE_Y=1 > perf_out
+$MPI_INSTALL_PATH/bin/mpirun -np 8 ./tealeaf_mpi_sycl OPS_CL_DEVICE=0 OPS_BLOCK_SIZE_X=256 OPS_BLOCK_SIZE_Y=1 > perf_out
 grep "Total Wall time" perf_out
 grep "PASSED" perf_out
 rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
@@ -207,7 +208,7 @@ rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
 rm -f tea.out
 
 if [[ -v CUDA_INSTALL_PATH ]]; then
-make IEEE=1 tealeaf_cuda tealeaf_mpi_cuda tealeaf_mpi_cuda_tiled 
+make IEEE=1 tealeaf_cuda tealeaf_mpi_cuda tealeaf_mpi_cuda_tiled
 #tealeaf_openacc tealeaf_mpi_openacc tealeaf_mpi_openacc_tiled
 
 echo '============> Running CUDA'
diff --git a/apps/c/adi/test.sh b/apps/c/adi/test.sh
index 1497ae848..abd77eca3 100755
--- a/apps/c/adi/test.sh
+++ b/apps/c/adi/test.sh
@@ -1,14 +1,24 @@
 #!/bin/bash
 set -e
-#<<COMMENT
 cd $OPS_INSTALL_PATH/c
+
 export SOURCE_INTEL=source_intel_2021.3_pythonenv
+
+#export AMOS=TRUE
+#export DMOS=TRUE
+#export TELOS=TRUE
+#export KOS=TRUE
+
+if [[ -v TELOS || -v KOS ]]; then
+
+#============================ Test with Intel Classic Compilers==========================================
+echo "Testing Intel classic complier based applications ---- "
+
 source ../../scripts/$SOURCE_INTEL
 
-#<<COMMENT
-#==== Build and copy Referance application from the TDMA Library ====
-#build lib first
-# TDMA_INSTALL_PATH=/path/to/tridsolver/scalar/build
+
+# Build and copy Referance application from the TDMA Library
+# build lib first
 cd $TDMA_INSTALL_PATH/../build
 rm -rf ./*
 cmake .. -DCUDA_cublas_LIBRARY=/opt/cuda/10.2.89/lib64/libcublas.so -DCMAKE_BUILD_TYPE=Release -DBUILD_FOR_CPU=ON -DBUILD_FOR_GPU=ON -DBUILD_FOR_SN=ON -DBUILD_FOR_MPI=ON -DCMAKE_INSTALL_PREFIX=$TDMA_INSTALL_PATH/../
@@ -22,7 +32,6 @@ cd $OPS_INSTALL_PATH/c
 make clean
 make IEEE=1
 
-#COMMENT
 
 #now build application
 cd $TDMA_INSTALL_PATH/../../apps/adi/build/
@@ -49,7 +58,7 @@ echo '============> Running Referance Solution adi_orig'
 ./adi_orig > perf_out
 rm perf_out
 
-#============================ Test adi with Intel Compilers==========================================================
+#============== Run ops ADI application ========================
 
 echo '============> Running DEV_SEQ'
 ./adi_dev_seq > perf_out
@@ -244,4 +253,8 @@ else
 fi
 
 rm -rf *.h5
-echo "All applications PASSED : Exiting Test Script"
+echo "All Intel classic complier based applications ---- PASSED"
+
+fi
+
+echo "---------- Exiting Test Script "
diff --git a/apps/c/lowdim_test/test.sh b/apps/c/lowdim_test/test.sh
index add68b90e..1bca6952b 100755
--- a/apps/c/lowdim_test/test.sh
+++ b/apps/c/lowdim_test/test.sh
@@ -1,54 +1,30 @@
 #!/bin/bash
 set -e
 cd ../../../ops/c
-<<COMMENT
-if [ -x "$(command -v enroot)" ]; then
-  cd -
-  enroot start --root --mount $OPS_INSTALL_PATH/../:/tmp/OPS --rw cuda112hip sh -c 'cd /tmp/OPS/apps/c/lowdim_test; ./test.sh'
-  grep "PASSED" perf_out
-  rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
-  rm perf_out
-  echo "All HIP complied applications PASSED"
-fi
 
-if [[ -v HIP_INSTALL_PATH ]]; then
-  source ../../scripts/$SOURCE_HIP
-  make -j -B
-  cd -
-  make clean
-  rm -f .generated
-  make lowdim_hip lowdim_mpi_hip -j
-     
-  echo '============> Running HIP'
-  ./lowdim_hip OPS_BLOCK_SIZE_X=64 OPS_BLOCK_SIZE_Y=4 > perf_out
-  #$HDF5_INSTALL_PATH/bin/h5diff output.h5 output_seq.h5
-  #rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi;
-  grep "PASSED" perf_out
-  rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
-  rm -rf perf_out output.h5
-  
-  echo '============> Running MPI+HIP'
-  mpirun --allow-run-as-root -np 2 ./lowdim_mpi_hip OPS_BLOCK_SIZE_X=64 OPS_BLOCK_SIZE_Y=4 > perf_out
-  #$HDF5_INSTALL_PATH/bin/h5diff output.h5 output_seq.h5
-  #rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi;
-  grep "PASSED" perf_out
-  rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
-  rm -rf perf_out output.h5
-  echo "All HIP complied applications PASSED : Moving no to Intel Compiler Tests " > perf_out
-  exit 0
-fi
-COMMENT
+export SOURCE_INTEL=source_intel_2021.3_pythonenv
+export SOURCE_PGI=source_pgi_nvhpc_23_pythonenv
+export SOURCE_INTEL_SYCL=source_intel_2021.3_sycl_pythonenv
+export SOURCE_AMD_HIP=source_amd_rocm-5.4.3_pythonenv
+
+#export AMOS=TRUE
+#export DMOS=TRUE
+#export TELOS=TRUE
+#export KOS=TRUE
+
+if [[ -v TELOS || -v KOS ]]; then
 
+#============================ Test with Intel Classic Compilers==========================================
+echo "Testing Intel classic complier based applications ---- "
 cd $OPS_INSTALL_PATH/c
 source ../../scripts/$SOURCE_INTEL
-make -j -B
-cd $OPS_INSTALL_PATH/../apps/c/lowdim_test
 make clean
-rm -f .generated
-make IEEE=1 -j
+make
 
+cd $OPS_INSTALL_PATH/../apps/c/lowdim_test
+make clean
+make IEEE=1
 
-#============================ Test lowdim with Intel Compilers ==========================================
 echo '============> Running SEQ'
 ./lowdim_dev_mpi > perf_out
 mv output.h5 output_seq.h5
@@ -88,6 +64,7 @@ grep "PASSED" perf_out
 rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
 rm -rf perf_out output.h5
 
+if [[ -v CUDA_INSTALL_PATH ]]; then
 echo '============> Running CUDA'
 ./lowdim_cuda OPS_BLOCK_SIZE_X=64 OPS_BLOCK_SIZE_Y=4 > perf_out
 #$HDF5_INSTALL_PATH/bin/h5diff output.h5 output_seq.h5
@@ -107,59 +84,71 @@ rm -rf perf_out output.h5
 #echo '============> Running MPI+CUDA with GPU-Direct'
 #MV2_USE_CUDA=1 $MPI_INSTALL_PATH/bin/mpirun -np 2 ./lowdim_mpi_cuda -gpudirect OPS_BLOCK_SIZE_X=64 OPS_BLOCK_SIZE_Y=4 > perf_out
 
-
 #grep "PASSED" perf_out
 #rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
 #rm -rf perf_out output.h5
 
-echo '============> Running OpenCL on CPU'
-./lowdim_opencl OPS_CL_DEVICE=0 OPS_BLOCK_SIZE_X=512 OPS_BLOCK_SIZE_Y=1 > perf_out
-#$HDF5_INSTALL_PATH/bin/h5diff output.h5 output_seq.h5
-#rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi;
-grep "PASSED" perf_out
-rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
-rm -rf perf_out output.h5
+fi
 
-echo '============> Running OpenCL on GPU'
-./lowdim_opencl OPS_CL_DEVICE=1 OPS_BLOCK_SIZE_X=32 OPS_BLOCK_SIZE_Y=4 > perf_out
-./lowdim_opencl OPS_CL_DEVICE=1 OPS_BLOCK_SIZE_X=32 OPS_BLOCK_SIZE_Y=4 > perf_out
+echo "All Intel classic complier based applications ---- PASSED"
+
+fi
+
+if [[ -v TELOS ]]; then
+
+#============================ Test with Intel SYCL Compilers==========================================
+echo "Testing Intel SYCL complier based applications ---- "
+cd $OPS_INSTALL_PATH/c
+source ../../scripts/$SOURCE_INTEL_SYCL
+
+make clean
+make
+cd $OPS_INSTALL_PATH/../apps/c/lowdim_test
+
+make clean
+make IEEE=1 lowdim_sycl lowdim_mpi_sycl lowdim_mpi_sycl_tiled
+
+echo '============> Running SYCL on CPU'
+./lowdim_sycl OPS_CL_DEVICE=0 OPS_BLOCK_SIZE_X=512 OPS_BLOCK_SIZE_Y=1 > perf_out
 #$HDF5_INSTALL_PATH/bin/h5diff output.h5 output_seq.h5
 #rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi;
 grep "PASSED" perf_out
 rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
 rm -rf perf_out output.h5
 
-echo '============> Running MPI+OpenCL on CPU'
-$MPI_INSTALL_PATH/bin/mpirun -np 20 ./lowdim_mpi_opencl OPS_CL_DEVICE=0 OPS_BLOCK_SIZE_X=256 OPS_BLOCK_SIZE_Y=1 > perf_out
-$MPI_INSTALL_PATH/bin/mpirun -np 20 ./lowdim_mpi_opencl OPS_CL_DEVICE=0 OPS_BLOCK_SIZE_X=256 OPS_BLOCK_SIZE_Y=1 > perf_out
+echo '============> Running MPI+SYCL on CPU'
+$MPI_INSTALL_PATH/bin/mpirun -np 12 ./lowdim_mpi_sycl OPS_CL_DEVICE=0 OPS_BLOCK_SIZE_X=256 OPS_BLOCK_SIZE_Y=1 > perf_out
 #$HDF5_INSTALL_PATH/bin/h5diff output.h5 output_seq.h5
 #rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi;
 grep "PASSED" perf_out
 rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
 rm -rf perf_out output.h5
 
-echo '============> Running MPI+OpenCL on GPU'
-$MPI_INSTALL_PATH/bin/mpirun -np 2 ./lowdim_mpi_opencl OPS_CL_DEVICE=1 OPS_BLOCK_SIZE_X=32 OPS_BLOCK_SIZE_Y=4 > perf_out
-$MPI_INSTALL_PATH/bin/mpirun -np 2 ./lowdim_mpi_opencl OPS_CL_DEVICE=1 OPS_BLOCK_SIZE_X=32 OPS_BLOCK_SIZE_Y=4 > perf_out
+echo '============> Running MPI+SYCL+Tiled on CPU'
+$MPI_INSTALL_PATH/bin/mpirun -np 2 ./lowdim_mpi_sycl_tiled OPS_CL_DEVICE=1 OPS_BLOCK_SIZE_X=32 OPS_BLOCK_SIZE_Y=4 > perf_out
 #$HDF5_INSTALL_PATH/bin/h5diff output.h5 output_seq.h5
 #rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi;
 grep "PASSED" perf_out
 rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
 rm -rf perf_out output.h5
 
-echo "All Intel complied applications PASSED : Moving no to PGI Compiler Tests "
+echo "All Intel SYCL complier based applications ---- PASSED"
 
+fi
 
-cd -
-source ../../scripts/$SOURCE_PGI
+if [[ -v TELOS ]]; then
 
+#============================ Test with PGI Compilers==========================================
+echo "Testing PGI/NVHPC complier based applications ---- "
+cd $OPS_INSTALL_PATH/c
+source ../../scripts/$SOURCE_PGI
 make clean
-make -j
-cd -
+#make -j
+make
+cd $OPS_INSTALL_PATH/../apps/c/lowdim_test
 make clean
-make -j
+make IEEE=1
 
-#============================ Test lowdim with PGI Compilers ==========================================
 echo '============> Running OpenMP'
 KMP_AFFINITY=compact OMP_NUM_THREADS=12 ./lowdim_openmp > perf_out
 #$HDF5_INSTALL_PATH/bin/h5diff output.h5 output_seq.h5
@@ -192,6 +181,7 @@ grep "PASSED" perf_out
 rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
 rm -rf perf_out output.h5
 
+if [[ -v CUDA_INSTALL_PATH ]]; then
 echo '============> Running CUDA'
 ./lowdim_cuda OPS_BLOCK_SIZE_X=64 OPS_BLOCK_SIZE_Y=4 > perf_out
 #$HDF5_INSTALL_PATH/bin/h5diff output.h5 output_seq.h5
@@ -223,48 +213,59 @@ rm -rf perf_out output.h5
 #grep "PASSED" perf_out
 #rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
 #rm -rf perf_out output.h5
+fi
 
-echo '============> Running OpenCL on GPU'
-./lowdim_opencl OPS_CL_DEVICE=1 OPS_BLOCK_SIZE_X=32 OPS_BLOCK_SIZE_Y=4 > perf_out
-./lowdim_opencl OPS_CL_DEVICE=1 OPS_BLOCK_SIZE_X=32 OPS_BLOCK_SIZE_Y=4 > perf_out
+echo '============> Running OMPOFFLOAD'
+./lowdim_ompoffload OPS_BLOCK_SIZE_X=32 OPS_BLOCK_SIZE_Y=4 > perf_out
 #$HDF5_INSTALL_PATH/bin/h5diff output.h5 output_seq.h5
 #rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi;
 grep "PASSED" perf_out
 rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
 rm -rf perf_out output.h5
-
-#echo '============> Running MPI+OpenCL on CPU'
-#$MPI_INSTALL_PATH/bin/mpirun -np 20 ./lowdim_mpi_opencl OPS_CL_DEVICE=0 OPS_BLOCK_SIZE_X=256 OPS_BLOCK_SIZE_Y=1 > perf_out
-#$MPI_INSTALL_PATH/bin/mpirun -np 20 ./lowdim_mpi_opencl OPS_CL_DEVICE=0 OPS_BLOCK_SIZE_X=256 OPS_BLOCK_SIZE_Y=1 > perf_out
-
-
-#grep "PASSED" perf_out
-#rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
-#rm -rf perf_out output.h5
-
-echo '============> Running MPI+OpenCL on GPU'
-$MPI_INSTALL_PATH/bin/mpirun -np 2 ./lowdim_mpi_opencl OPS_CL_DEVICE=1 OPS_BLOCK_SIZE_X=32 OPS_BLOCK_SIZE_Y=4 > perf_out
-$MPI_INSTALL_PATH/bin/mpirun -np 2 ./lowdim_mpi_opencl OPS_CL_DEVICE=1 OPS_BLOCK_SIZE_X=32 OPS_BLOCK_SIZE_Y=4 > perf_out
+#COMMENT
+echo '============> Running MPI+OMPOFFLOAD'
+$MPI_INSTALL_PATH/bin/mpirun -np 2 ./lowdim_mpi_ompoffload numawrap2 OPS_BLOCK_SIZE_X=32 OPS_BLOCK_SIZE_Y=4 > perf_out
 #$HDF5_INSTALL_PATH/bin/h5diff output.h5 output_seq.h5
 #rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi;
 grep "PASSED" perf_out
 rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
 rm -rf perf_out output.h5
+echo "All PGI complied applications PASSED : Exiting Test Script "
+
+echo "All PGI complier based applications ---- PASSED"
+
+fi
+
+if [[ -v AMOS ]]; then
+
+echo "Testing AMD HIP complier based applications ---- "
+cd $OPS_INSTALL_PATH/c
+source ../../scripts/$SOURCE_AMD_HIP
+make clean
+make
+cd $OPS_INSTALL_PATH/../apps/c/lowdim_test
+
+make clean
+make lowdim_hip lowdim_mpi_hip
 
-echo '============> Running OpenACC'
-./lowdim_openacc OPS_BLOCK_SIZE_X=32 OPS_BLOCK_SIZE_Y=4 > perf_out
+echo '============> Running HIP'
+./lowdim_hip OPS_BLOCK_SIZE_X=64 OPS_BLOCK_SIZE_Y=4 > perf_out
 #$HDF5_INSTALL_PATH/bin/h5diff output.h5 output_seq.h5
 #rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi;
 grep "PASSED" perf_out
 rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
 rm -rf perf_out output.h5
-#COMMENT
-echo '============> Running MPI+OpenACC'
-$MPI_INSTALL_PATH/bin/mpirun -np 2 ./lowdim_mpi_openacc numawrap2 OPS_BLOCK_SIZE_X=32 OPS_BLOCK_SIZE_Y=4 > perf_out
+
+echo '============> Running MPI+HIP'
+mpirun --allow-run-as-root -np 2 ./lowdim_mpi_hip OPS_BLOCK_SIZE_X=64 OPS_BLOCK_SIZE_Y=4 > perf_out
 #$HDF5_INSTALL_PATH/bin/h5diff output.h5 output_seq.h5
 #rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi;
 grep "PASSED" perf_out
 rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
 rm -rf perf_out output.h5
-echo "All PGI complied applications PASSED : Exiting Test Script "
 
+echo "All AMD HIP complier based applications ---- PASSED"
+
+fi
+
+echo "---------- Exiting Test Script "
\ No newline at end of file
diff --git a/apps/c/mb_shsgc/Max_datatransfer/test.sh b/apps/c/mb_shsgc/Max_datatransfer/test.sh
index f205f7c94..045b4c293 100755
--- a/apps/c/mb_shsgc/Max_datatransfer/test.sh
+++ b/apps/c/mb_shsgc/Max_datatransfer/test.sh
@@ -1,59 +1,29 @@
 #!/bin/bash
 set -e
 cd $OPS_INSTALL_PATH/c
-#f [ -x "$(command -v enroot)" ]; then
-# -----
-# Fails to compile with NVCC
-# -----
-#  cd -
-#  enroot start --root --mount $OPS_INSTALL_PATH/../:/tmp/OPS --rw cuda112hip sh -c 'cd /tmp/OPS/apps/c/mb_shsgc/Max_datatransfer; ./test.sh'
-#  grep "PASSED" perf_out
-#  rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
-#  rm perf_out
-#  echo "All HIP complied applications PASSED"
-#i
-
-<<COMMENT
-if [[ -v HIP_INSTALL_PATH ]]; then
-  source ../../scripts/$SOURCE_HIP
-  make -j -B
-  cd -
-  make clean
-  rm -f .generated
-  make shsgc_hip shsgc_mpi_hip -j
-   
-  echo '============> Running HIP'
-  ./shsgc_hip OPS_BLOCK_SIZE_X=64 OPS_BLOCK_SIZE_Y=4 > perf_out
-  grep "Pre shock error is:" perf_out
-  grep "Post shock error is:" perf_out
-  grep "Post shock Error is" perf_out
-  grep "Total Wall time" perf_out
-  grep -e "acceptable" -e "correct"  perf_out
-  rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
-  rm perf_out
-  
-  echo '============> Running MPI+HIP'
-  mpirun --allow-run-as-root -np 2 ./shsgc_mpi_hip OPS_BLOCK_SIZE_X=64 OPS_BLOCK_SIZE_Y=4 > perf_out
-  grep "Pre shock error is:" perf_out
-  grep "Post shock error is:" perf_out
-  grep "Post shock Error is" perf_out
-  grep "Total Wall time" perf_out
-  grep -e "acceptable" -e "correct"  perf_out
-  rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
-  rm perf_out
-  echo "All HIP complied applications PASSED : Moving no to Intel Compiler Tests" > perf_out
-  exit 0
-fi
-COMMENT
 
+export SOURCE_INTEL=source_intel_2021.3_pythonenv
+export SOURCE_PGI=source_pgi_nvhpc_23_pythonenv
+export SOURCE_INTEL_SYCL=source_intel_2021.3_sycl_pythonenv
+export SOURCE_AMD_HIP=source_amd_rocm-5.4.3_pythonenv
+
+#export AMOS=TRUE
+#export DMOS=TRUE
+#export TELOS=TRUE
+#export KOS=TRUE
+
+if [[ -v TELOS || -v KOS ]]; then
+
+#============================ Test with Intel Classic Compilers==========================================
+echo "Testing Intel classic complier based applications ---- "
 cd $OPS_INSTALL_PATH/c
 source ../../scripts/$SOURCE_INTEL
-make -j -B
-cd $OPS_INSTALL_PATH/../apps/c/mb_shsgc/Max_datatransfer
 make clean
-rm -f .generated
-make IEEE=1 -j
+make
 
+cd $OPS_INSTALL_PATH/../apps/c/mb_shsgc/Max_datatransfer
+make clean
+make IEEE=1
 
 
 #============================ Test SHSGC with Intel Compilers==========================================================
@@ -97,6 +67,7 @@ grep -e "acceptable" -e "correct"  perf_out
 rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
 rm perf_out
 
+if [[ -v CUDA_INSTALL_PATH ]]; then
 echo '============> Running CUDA'
 ./shsgc_cuda OPS_BLOCK_SIZE_X=64 OPS_BLOCK_SIZE_Y=4 > perf_out
 grep "Pre shock error is:" perf_out
@@ -126,22 +97,31 @@ rm perf_out
 #grep -e "acceptable" -e "correct"  perf_out
 #rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
 #rm perf_out
+fi
 
-<<COMMENT
-echo '============> Running OpenCL on CPU'
-./shsgc_opencl OPS_CL_DEVICE=0 OPS_BLOCK_SIZE_X=512 OPS_BLOCK_SIZE_Y=1 > perf_out
-grep "Pre shock error is:" perf_out
-grep "Post shock error is:" perf_out
-grep "Post shock Error is" perf_out
-grep "Total Wall time" perf_out
-grep -e "acceptable" -e "correct"  perf_out
-rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
-rm perf_out
-COMMENT
+#cleanup
+rm rhoin1.* x1.* rhoout1.*  rhoin1 rhoout1 x1
+
+echo "All Intel classic complier based applications ---- PASSED"
+
+fi
 
-echo '============> Running OpenCL on GPU'
-./shsgc_opencl OPS_CL_DEVICE=1 OPS_BLOCK_SIZE_X=32 OPS_BLOCK_SIZE_Y=4 > perf_out
-./shsgc_opencl OPS_CL_DEVICE=1 OPS_BLOCK_SIZE_X=32 OPS_BLOCK_SIZE_Y=4 > perf_out
+
+if [[ -v TELOS ]]; then
+	#============================ Test with Intel SYCL Compilers==========================================
+echo "Testing Intel SYCL complier based applications ---- "
+cd $OPS_INSTALL_PATH/c
+source ../../scripts/$SOURCE_INTEL_SYCL
+make clean
+make
+cd $OPS_INSTALL_PATH/../apps/c/mb_shsgc/Max_datatransfer
+
+make clean
+#make IEEE=1 -j
+make IEEE=1 shsgc_sycl shsgc_mpi_sycl shsgc_mpi_sycl_tiled
+
+echo '============> Running SYCL on CPU'
+./shsgc_sycl OPS_CL_DEVICE=0 OPS_BLOCK_SIZE_X=32 OPS_BLOCK_SIZE_Y=4 > perf_out
 grep "Pre shock error is:" perf_out
 grep "Post shock error is:" perf_out
 grep "Post shock Error is" perf_out
@@ -150,10 +130,8 @@ grep -e "acceptable" -e "correct"  perf_out
 rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
 rm perf_out
 
-<<COMMENT
-echo '============> Running MPI+OpenCL on CPU'
-$MPI_INSTALL_PATH/bin/mpirun -np 20 ./shsgc_mpi_opencl OPS_CL_DEVICE=0 OPS_BLOCK_SIZE_X=256 OPS_BLOCK_SIZE_Y=1 > perf_out
-$MPI_INSTALL_PATH/bin/mpirun -np 20 ./shsgc_mpi_opencl OPS_CL_DEVICE=0 OPS_BLOCK_SIZE_X=256 OPS_BLOCK_SIZE_Y=1 > perf_out
+echo '============> Running MPI+SYCL on CPU'
+$MPI_INSTALL_PATH/bin/mpirun -np 12 ./shsgc_mpi_sycl OPS_CL_DEVICE=0 OPS_BLOCK_SIZE_X=256 OPS_BLOCK_SIZE_Y=1 > perf_out
 grep "Pre shock error is:" perf_out
 grep "Post shock error is:" perf_out
 grep "Post shock Error is" perf_out
@@ -161,11 +139,9 @@ grep "Total Wall time" perf_out
 grep -e "acceptable" -e "correct"  perf_out
 rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
 rm perf_out
-COMMENT
 
-echo '============> Running MPI+OpenCL on GPU'
-$MPI_INSTALL_PATH/bin/mpirun -np 2 ./shsgc_mpi_opencl OPS_CL_DEVICE=1 OPS_BLOCK_SIZE_X=32 OPS_BLOCK_SIZE_Y=4 > perf_out
-$MPI_INSTALL_PATH/bin/mpirun -np 2 ./shsgc_mpi_opencl OPS_CL_DEVICE=1 OPS_BLOCK_SIZE_X=32 OPS_BLOCK_SIZE_Y=4 > perf_out
+echo '============> Running MPI+SYCL Tiled on CPU'
+$MPI_INSTALL_PATH/bin/mpirun -np 2 ./shsgc_mpi_sycl_tiled OPS_CL_DEVICE=1 OPS_BLOCK_SIZE_X=32 OPS_BLOCK_SIZE_Y=4 > perf_out
 grep "Pre shock error is:" perf_out
 grep "Post shock error is:" perf_out
 grep "Post shock Error is" perf_out
@@ -177,16 +153,24 @@ rm perf_out
 #cleanup
 rm rhoin1.* x1.* rhoout1.*  rhoin1 rhoout1 x1
 
-cd -
-source ../../scripts/$SOURCE_PGI
+echo "All Intel SYCL complier based applications ---- PASSED"
 
+fi
+
+
+if [[ -v TELOS ]]; then
+
+#============================ Test with PGI Compilers==========================================
+echo "Testing PGI/NVHPC complier based applications ---- "
+cd $OPS_INSTALL_PATH/c
+source ../../scripts/$SOURCE_PGI
 make clean
-make 
-cd -
-make clean
+#make -j
 make
+cd $OPS_INSTALL_PATH/../apps/c/mb_shsgc/Max_datatransfer
+make clean
+make IEEE=1
 
-#============================ Test SHSGC with PGI Compilers==========================================================
 echo '============> Running OpenMP'
 KMP_AFFINITY=compact OMP_NUM_THREADS=10 ./shsgc_openmp > perf_out
 grep "Pre shock error is:" perf_out
@@ -227,6 +211,7 @@ grep -e "acceptable" -e "correct"  perf_out
 rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
 rm perf_out
 
+if [[ -v CUDA_INSTALL_PATH ]]; then
 echo '============> Running CUDA'
 ./shsgc_cuda OPS_BLOCK_SIZE_X=64 OPS_BLOCK_SIZE_Y=4 > perf_out
 grep "Pre shock error is:" perf_out
@@ -246,6 +231,17 @@ grep "Total Wall time" perf_out
 grep -e "acceptable" -e "correct"  perf_out
 rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
 rm perf_out
+
+echo '============> Running MPI+CUDA+Tiled'
+$MPI_INSTALL_PATH/bin/mpirun -np 2 ./shsgc_mpi_cuda_tiled OPS_BLOCK_SIZE_X=64 OPS_BLOCK_SIZE_Y=4 > perf_out
+grep "Pre shock error is:" perf_out
+grep "Post shock error is:" perf_out
+grep "Post shock Error is" perf_out
+grep "Total Wall time" perf_out
+grep -e "acceptable" -e "correct"  perf_out
+rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
+rm perf_out
+
 #echo '============> Running MPI+CUDA with GPU-Direct'
 #MV2_USE_CUDA=1 $MPI_INSTALL_PATH/bin/mpirun -np 2 ./shsgc_mpi_cuda -gpudirect OPS_BLOCK_SIZE_X=64 OPS_BLOCK_SIZE_Y=4 > perf_out
 #grep "Pre shock error is:" perf_out
@@ -256,19 +252,10 @@ rm perf_out
 #rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
 #rm perf_out
 
-#echo '============> Running OpenCL on CPU'
-#./shsgc_opencl OPS_CL_DEVICE=0 OPS_BLOCK_SIZE_X=512 OPS_BLOCK_SIZE_Y=1 > perf_out
-#grep "Pre shock error is:" perf_out
-#grep "Post shock error is:" perf_out
-#grep "Post shock Error is" perf_out
-#grep "Total Wall time" perf_out
-#grep -e "acceptable" -e "correct"  perf_out
-#rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
-#rm perf_out
+fi
 
-echo '============> Running OpenCL on GPU'
-./shsgc_opencl OPS_CL_DEVICE=1 OPS_BLOCK_SIZE_X=32 OPS_BLOCK_SIZE_Y=4 > perf_out
-./shsgc_opencl OPS_CL_DEVICE=1 OPS_BLOCK_SIZE_X=32 OPS_BLOCK_SIZE_Y=4 > perf_out
+echo '============> Running OMPOFFLOAD'
+./shsgc_offload OPS_BLOCK_SIZE_X=64 OPS_BLOCK_SIZE_Y=4 > perf_out
 grep "Pre shock error is:" perf_out
 grep "Post shock error is:" perf_out
 grep "Post shock Error is" perf_out
@@ -277,29 +264,37 @@ grep -e "acceptable" -e "correct"  perf_out
 rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
 rm perf_out
 
-#echo '============> Running MPI+OpenCL on CPU'
-#$MPI_INSTALL_PATH/bin/mpirun -np 20 ./shsgc_mpi_opencl OPS_CL_DEVICE=0 OPS_BLOCK_SIZE_X=256 OPS_BLOCK_SIZE_Y=1 > perf_out
-#$MPI_INSTALL_PATH/bin/mpirun -np 20 ./shsgc_mpi_opencl OPS_CL_DEVICE=0 OPS_BLOCK_SIZE_X=256 OPS_BLOCK_SIZE_Y=1 > perf_out
-#grep "Pre shock error is:" perf_out
-#grep "Post shock error is:" perf_out
-#grep "Post shock Error is" perf_out
-#grep "Total Wall time" perf_out
-#grep -e "acceptable" -e "correct"  perf_out
-#rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
-#rm perf_out
-
-echo '============> Running MPI+OpenCL on GPU'
-$MPI_INSTALL_PATH/bin/mpirun -np 2 ./shsgc_mpi_opencl OPS_CL_DEVICE=1 OPS_BLOCK_SIZE_X=32 OPS_BLOCK_SIZE_Y=4 > perf_out
-$MPI_INSTALL_PATH/bin/mpirun -np 2 ./shsgc_mpi_opencl OPS_CL_DEVICE=1 OPS_BLOCK_SIZE_X=32 OPS_BLOCK_SIZE_Y=4 > perf_out
+echo '============> Running MPI+OMPOFFLOAD'
+$MPI_INSTALL_PATH/bin/mpirun -np 2 ./shsgc_mpi_offload OPS_BLOCK_SIZE_X=64 OPS_BLOCK_SIZE_Y=4 > perf_out
 grep "Pre shock error is:" perf_out
 grep "Post shock error is:" perf_out
 grep "Post shock Error is" perf_out
 grep "Total Wall time" perf_out
 grep -e "acceptable" -e "correct"  perf_out
 rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
+rm perf_out
+
+#cleanup
+rm rhoin1.* x1.* rhoout1.*  rhoin1 rhoout1 x1
+
+echo "All PGI complier based applications ---- PASSED"
+
+fi
+
+if [[ -v AMOS ]]; then
+
+echo "Testing AMD HIP complier based applications ---- "
+cd $OPS_INSTALL_PATH/c
+source ../../scripts/$SOURCE_AMD_HIP
+make clean
+make
+
+cd $OPS_INSTALL_PATH/../apps/c/mb_shsgc/Max_datatransfer
+make clean
+make IEEE=1 shsgc_hip shsgc_mpi_hip
 
-echo '============> Running OpenACC'
-./shsgc_openacc OPS_BLOCK_SIZE_X=64 OPS_BLOCK_SIZE_Y=4 > perf_out
+echo '============> Running HIP'
+./shsgc_hip OPS_BLOCK_SIZE_X=64 OPS_BLOCK_SIZE_Y=4 > perf_out
 grep "Pre shock error is:" perf_out
 grep "Post shock error is:" perf_out
 grep "Post shock Error is" perf_out
@@ -308,8 +303,8 @@ grep -e "acceptable" -e "correct"  perf_out
 rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
 rm perf_out
 
-echo '============> Running MPI+OpenACC'
-$MPI_INSTALL_PATH/bin/mpirun -np 2 ./shsgc_mpi_openacc OPS_BLOCK_SIZE_X=64 OPS_BLOCK_SIZE_Y=4 > perf_out
+echo '============> Running MPI+HIP'
+mpirun --allow-run-as-root -np 2 ./shsgc_mpi_hip OPS_BLOCK_SIZE_X=64 OPS_BLOCK_SIZE_Y=4 > perf_out
 grep "Pre shock error is:" perf_out
 grep "Post shock error is:" perf_out
 grep "Post shock Error is" perf_out
@@ -318,6 +313,11 @@ grep -e "acceptable" -e "correct"  perf_out
 rc=$?; if [[ $rc != 0 ]]; then echo "TEST FAILED";exit $rc; fi
 rm perf_out
 
+echo "All AMD HIP complier based applications ---- PASSED"
+
 #cleanup
 rm rhoin1.* x1.* rhoout1.*  rhoin1 rhoout1 x1
-echo "All PGI Tests Passed. Exiting test script"
+
+fi
+
+echo "---------- Exiting Test Script "