More cleanup of OpenMP accelerator code

OpenMP/StreamTriad/CMakeLists.txt

@@ -45,6 +45,11 @@ add_executable(StreamTriad_par3 StreamTriad_par3.c timer.c timer.h)
 set_target_properties(StreamTriad_par3 PROPERTIES COMPILE_FLAGS ${OpenMPAccel_C_FLAGS})
 set_target_properties(StreamTriad_par3 PROPERTIES LINK_FLAGS "${OpenMPAccel_C_FLAGS}")
 
+# Adds build target of stream_triad_par4 with source code files
+add_executable(StreamTriad_par4 StreamTriad_par4.c timer.c timer.h)
+set_target_properties(StreamTriad_par4 PROPERTIES COMPILE_FLAGS ${OpenMPAccel_C_FLAGS})
+set_target_properties(StreamTriad_par4 PROPERTIES LINK_FLAGS "${OpenMPAccel_C_FLAGS}")
+
 # Cleanup
 add_custom_target(distclean COMMAND rm -rf CMakeCache.txt CMakeFiles
                   Makefile cmake_install.cmake StreamTriad.dSYM ipo_out.optrpt)

OpenMP/StreamTriad/StreamTriad_par1.c

@@ -33,18 +33,8 @@ int main(int argc, char *argv[]){
       time_sum += cpu_timer_stop(tstart);
    }
 
-
    printf("Average runtime for stream triad loop is %lf msecs\n", time_sum/ntimes);
 
-   for (int k=0; k<ntimes; k++){
-      cpu_timer_start(&tstart);
-      // stream triad function call
-      StreamTriad(c, a, b, scalar, nsize);
-      time_sum += cpu_timer_stop(tstart);
-   }
-
-   printf("Average runtime for stream triad function call is %lf msecs\n", time_sum/ntimes);
-
    free(a);
    free(b);
    free(c);

OpenMP/StreamTriad/StreamTriad_par2.c

@@ -12,34 +12,35 @@ int main(int argc, char *argv[]){
    double *b = malloc(nsize * sizeof(double));
    double *c = malloc(nsize * sizeof(double));
 
-#pragma omp target enter data map(to:a[0:nsize], b[0:nsize], c[0:nsize])
-
    struct timespec tstart;
    // initializing data and arrays
    double scalar = 3.0, time_sum = 0.0;
+
+#pragma omp target data map(to:a[0:nsize], b[0:nsize], c[0:nsize])
+   {
+
 #pragma omp target
 #pragma omp teams distribute parallel for
-   for (int i=0; i<nsize; i++) {
-      a[i] = 1.0;
-      b[i] = 2.0;
-   }
+      for (int i=0; i<nsize; i++) {
+         a[i] = 1.0;
+         b[i] = 2.0;
+      }
 
-   for (int k=0; k<ntimes; k++){
-      cpu_timer_start(&tstart);
-      // stream triad loop 
+      for (int k=0; k<ntimes; k++){
+         cpu_timer_start(&tstart);
+         // stream triad loop 
 #pragma omp target
 #pragma omp teams distribute parallel for
-      for (int i=0; i<nsize; i++){
-         c[i] = a[i] + scalar*b[i];
+         for (int i=0; i<nsize; i++){
+            c[i] = a[i] + scalar*b[i];
+         }
+         time_sum += cpu_timer_stop(tstart);
       }
-      time_sum += cpu_timer_stop(tstart);
-   }
 
+   } // #pragma omp target data map(from:c[0:nsize])
 
    printf("Average runtime for stream triad loop is %lf msecs\n", time_sum/ntimes);
 
-#pragma omp target exit data map(from:c[0:nsize])
-
    free(a);
    free(b);
    free(c);

OpenMP/StreamTriad/StreamTriad_par3.c

@@ -8,12 +8,11 @@ void StreamTriad(double *restrict c, double *restrict a, double *restrict b, dou
 int main(int argc, char *argv[]){
 
    int nsize = 20000000, ntimes=16;
-   //double *a = malloc(nsize * sizeof(double));
-   //double *b = malloc(nsize * sizeof(double));
-   //double *c = malloc(nsize * sizeof(double));
-   double *a, *b, *c;
+   double *a = malloc(nsize * sizeof(double));
+   double *b = malloc(nsize * sizeof(double));
+   double *c = malloc(nsize * sizeof(double));
 
-#pragma omp target enter data map(alloc:a[0:nsize], b[0:nsize], c[0:nsize])
+#pragma omp target enter data map(to:a[0:nsize], b[0:nsize], c[0:nsize])
 
    struct timespec tstart;
    // initializing data and arrays
@@ -36,14 +35,13 @@ int main(int argc, char *argv[]){
       time_sum += cpu_timer_stop(tstart);
    }
 
-
    printf("Average runtime for stream triad loop is %lf msecs\n", time_sum/ntimes);
 
-#pragma omp target exit data map(delete:a[0:nsize], b[0:nsize], c[0:nsize])
+#pragma omp target exit data map(from:a[0:nsize], b[0:nsize], c[0:nsize])
 
-   //free(a);
-   //free(b);
-   //free(c);
+   free(a);
+   free(b);
+   free(c);
 
    return(0);
 }

OpenMP/StreamTriad/StreamTriad_par4.c

@@ -0,0 +1,47 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <time.h>
+#include "timer.h"
+
+void StreamTriad(double *restrict c, double *restrict a, double *restrict b, double scalar, int nsize);
+
+int main(int argc, char *argv[]){
+
+   int nsize = 20000000, ntimes=16;
+   double *a, *b, *c;
+
+#pragma omp target enter data map(alloc:a[0:nsize], b[0:nsize], c[0:nsize])
+
+   struct timespec tstart;
+   // initializing data and arrays
+   double scalar = 3.0, time_sum = 0.0;
+#pragma omp target
+#pragma omp teams distribute parallel for
+   for (int i=0; i<nsize; i++) {
+      a[i] = 1.0;
+      b[i] = 2.0;
+   }
+
+   for (int k=0; k<ntimes; k++){
+      cpu_timer_start(&tstart);
+      // stream triad loop 
+#pragma omp target
+#pragma omp teams distribute parallel for
+      for (int i=0; i<nsize; i++){
+         c[i] = a[i] + scalar*b[i];
+      }
+      time_sum += cpu_timer_stop(tstart);
+   }
+
+   printf("Average runtime for stream triad loop is %lf msecs\n", time_sum/ntimes);
+
+#pragma omp target exit data map(delete:a[0:nsize], b[0:nsize], c[0:nsize])
+
+   return(0);
+}
+
+void StreamTriad(double *restrict c, double *restrict a, double *restrict b, double scalar, int nsize){
+   for (int i=0; i<nsize; i++){
+      c[i] = a[i] + scalar*b[i];
+   }
+}