baseline performance is collected

ocn/gm_compute_Bolus_velocity/Makefile

@@ -36,7 +36,7 @@ _PREPROCFLAG :=
 # 1<= verbosity <= 3
 VERBOSITY ?= 1
 # repeat >= 1
-REPEAT ?= 1
+REPEAT ?= 1 #2
 # skip sum check during data loading
 SKIP_SUMCHECK ?=
 
@@ -69,7 +69,7 @@ else ifeq (${COMP}, crayftn)
 else ifeq (${COMP}, pgfortran)
 
     FC_0 := pgfortran
-	#FC_FLAGS_SET_0 := -i4 -time -Mstack_arrays -Mextend -byteswapio -Mflushz -Kieee -Mallocatable=03 -O2 -Mpreprocess
+	#FC_FLAGS_SET_0 := -gopt -i4 -time -Mstack_arrays -Mextend -byteswapio -Mflushz -Kieee -Mallocatable=03 -O2 -Mpreprocess
 	FC_FLAGS_SET_0 := -gopt -i4 -time -Mstack_arrays -Mextend -byteswapio -Mflushz -Kieee -Mallocatable=03 -O2 -Mpreprocess -ta=tesla:debug,lineinfo -Minfo=all,ccff
 	_PREPROCFLAG := -Mpreprocess
 
@@ -170,11 +170,10 @@ ncu: ${OUTDIR}/${APP}.compute.ncu-rep
 	@echo "Output files are in ${OUTDIR}"
 
 ${OUTDIR}/${APP}.compute.ncu-rep: ${OUTDIR}/${APP}
-	#jsrun -n1 -c1 -g1 -a1 --smpiargs="-disable_gpu_hooks" -- ncu --target-processes=all -c 1500 --set=full --force-overwrite -o ${OUTDIR}/${APP}.compute ${OUTDIR}/${APP}
 	jsrun -n1 -c1 -g1 -a1 --smpiargs="-disable_gpu_hooks" -- ncu --target-processes=all --set=full --force-overwrite -o ${OUTDIR}/${APP}.compute ${OUTDIR}/${APP}
 
 ${OUTDIR}/${APP}.systems.qdrep: ${OUTDIR}/${APP}
-	jsrun -n1 -c1 -g1 -a1 -- nsys profile -o ${OUTDIR}/${APP}.systems.qdrep -f true -t cuda,osrt,openacc ${OUTDIR}/${APP}
+	jsrun -n1 -c1 -g1 -a1 -- nsys profile -o ${OUTDIR}/${APP}.systems.qdrep -f true -t cuda,osrt,openacc ${OUTDIR}/${APP} || true
 
 ${OUTDIR}/${APP}: build
 
@@ -185,7 +184,7 @@ run: build
 	${_MPIRUN} ./kernel.exe
 
 mpas_ocn_gm.o: mpas_ocn_gm.f90 mpas_ocn_constants.o mpas_constants.o mpas_kind_types.o kgen_utils.o tprof_mod.o
-	${FC_0} ${FC_FLAGS_SET_0} ${_PREPROCFLAG} -DMAX_TOL=${MAX_TOL} -DVERBOSITY=${VERBOSITY} -DNUM_REPEAT=${REPEAT} -c -o $@ $<
+	${FC_0} ${FC_FLAGS_SET_0} ${_PREPROCFLAG} -DMAX_TOL=${MAX_TOL} -DVERBOSITY=${VERBOSITY} -DNUM_REPEAT=${REPEAT} -c -o $@ $^
 
 mpas_ocn_constants.o: mpas_ocn_constants.f90 kgen_utils.o tprof_mod.o mpas_kind_types.o
 	${FC_0} ${FC_FLAGS_SET_0} -c -o $@ $<

ocn/gm_compute_Bolus_velocity/kernel_driver.F90

@@ -138,14 +138,14 @@
             IF (kgen_case_count == 0) THEN 
                 WRITE (*, *) "No data file is verified." 
             ELSE 
-                WRITE (*, "(4X, A36, A1, I6)") "Total number of verification cases   ", ":", kgen_case_count 
-                WRITE (*, "(4X, A36, A1, I6)") "Number of verification-passed cases ", ":", kgen_count_verified 
-                WRITE (*, *) "" 
-                IF (kgen_case_count == kgen_count_verified) THEN 
-                    WRITE (*, "(4X,A)") "kernel: ocn_gm_compute_Bolus_velocity: PASSED verification" 
-                ELSE 
-                    WRITE (*, "(4X,A)") "kernel: ocn_gm_compute_Bolus_velocity: FAILED verification" 
-                END IF   
+!                WRITE (*, "(4X, A36, A1, I6)") "Total number of verification cases   ", ":", kgen_case_count 
+!                WRITE (*, "(4X, A36, A1, I6)") "Number of verification-passed cases ", ":", kgen_count_verified 
+!                WRITE (*, *) "" 
+!                IF (kgen_case_count == kgen_count_verified) THEN 
+!                    WRITE (*, "(4X,A)") "kernel: ocn_gm_compute_Bolus_velocity: PASSED verification" 
+!                ELSE 
+!                    WRITE (*, "(4X,A)") "kernel: ocn_gm_compute_Bolus_velocity: FAILED verification" 
+!                END IF   
                 WRITE (*, *) "" 
                 WRITE (*, "(4X,A19,I3)") "number of processes: ", mpisize 
                 WRITE (*, *) "" 

ocn/gm_compute_Bolus_velocity/timing.log

+
+CPU
+
+****************************************************
+    kernel execution summary: ocn_gm_compute_Bolus_velocity
+****************************************************
+ 
+    number of processes  1
+ 
+    Average call time (usec):  0.377E+05
+    Minimum call time (usec):  0.377E+05
+    Maximum call time (usec):  0.377E+05
+****************************************************
+
+GPU - baseline
+
+****************************************************
+    kernel execution summary: ocn_gm_compute_Bolus_velocity
+****************************************************
+ 
+    number of processes  1
+ 
+    Average call time (usec):  0.388E+05
+    Minimum call time (usec):  0.388E+05
+    Maximum call time (usec):  0.388E+05
+****************************************************