tes

src/radixsort_parallel.cu

-#include <cuda.h>
-#include <cuda_runtime.h>
-#include "radix_sort_parallel.h"
+#include <stdio.h>
+#include <stdlib.h>
+// #include <omp.h>
+#include <sys/time.h>
+
+void generateArray(int arr[], int n,int seed);
+void radixsort(int arr[], int n);
+void print(int arr[], int n);
+
+int main(int argc, char *argv[]) {
+   struct timeval stop, start;
+   // int thread_count = 5;
+   int array_size = strtol(argv[1], NULL, 10);
+    //printf("NUM THREADS : %d\n", omp_get_num_threads());
+   int *arr,*d_arr;
+   arr = (int*)malloc(sizeof(int)* array_size);
+   cudaMalloc((void **)&d_arr, sizeof(int) * array_size);
+//   cudaMemcpy(d_arr, arr, sizeof(int) * array_size, cudaMemcpyHostToDevice);
+   int seed = 13516127;
+   generateArray(arr,array_size,seed);
+   cudaMemcpy(d_arr, arr, sizeof(int) * array_size, cudaMemcpyHostToDevice);
+   gettimeofday(&start, NULL);
+   radixsort(arr, array_size);
+   gettimeofday(&stop, NULL);
+     print(arr,array_size);
+   printf("Sorting selesai\n");
+   printf("took %lu microsecond for parallel radix sort\n", ((stop.tv_sec - start.tv_sec)*1000000)+(stop.tv_usec - start.tv_usec));
+  
+   cudaFree(d_arr);
+   free(arr);
+    return 0;
+}
 
 __global__
 void getMax(int arr[], int n, int max[]) {
@@ -29,13 +58,23 @@ int temp = arr[0];
         }
     
     }
-} 
+}
+// __global__
+void generateArray(int arr[], int n,int seed){
+   int *output;
+   output = (int*)malloc(sizeof(int) * n);
+   srand(seed);
+   for(long i = 0; i < n; i++) {
+        output[i] = (int)rand();
+   }
+
+   for(long i = 0; i < n; i++) {
+      arr[i] = output[i];
+   }
 
+}
 __global__
 void countSort(int arr[], int n, int digit) {
-   // int my_rank = omp_get_thread_num();
-   // int thread_count = omp_get_num_threads();
-    //printf("NUMBER OF THREAD NYA BOY %d\n",thread_count);
     int *output;
     output = (int*)malloc(sizeof(int)*n);
   long i;
@@ -53,8 +92,7 @@ void countSort(int arr[], int n, int digit) {
      arr[i] = output[i];
 }
 
-void radix_sort(int arr[], int n) 
-{ 
+void radixsort(int arr[], int n) {
   int  *max,*d_max,*d_arr;
   max = (int*)malloc(sizeof(int)* 1);
   cudaMalloc((void **)&d_max, sizeof(int) * 1);
@@ -67,55 +105,13 @@ void radix_sort(int arr[], int n)
       countSort<<<1,1>>>(d_arr, n, digit);
    }
   cudaMemcpy(arr,d_arr, sizeof(int) *n, cudaMemcpyDeviceToHost);
-} 
-
-void print(int arr[], int n) 
-{ 
-    for (int i = 0; i < n; i++) 
-        printf("%d: %d\n",i, arr[i]);
-} 
 
-void rng(int *arr, int n) {
-    int seed = 13516013;
-    srand(seed);
-    for(long i = 0; i < n; i++) {
-        arr[i] = (int)rand();
-    }
 }
 
-int main(int argc, char *argv[]) {
-    int N;
-    int *arr;
-    int *d_arr;
-
-    if (argc == 2) {
-        N = strtol(argv[1], NULL, 10);
-    } else {
-        printf("ERROR: ./radix_sort <array_length>\n");
-        return 1;
-    }
-
-    // Allocate host memory
-    arr = (int*)malloc(N * sizeof(int));
-
-    // Initialize host memory
-    rng(arr,N);
-
-    // Allocate device memory
-    cudaMalloc((void**)&d_arr, N * sizeof(int));
-
-    // Transfer data from host to device memory
-    cudaMemcpy(d_arr, arr, N * sizeof(int), cudaMemcpyHostToDevice);
-
-    clock_t begin = clock();
-    radix_sort(arr, N);
-    clock_t end = clock();
-    double time = (double)(end - begin) * 1000 / CLOCKS_PER_SEC;
-    print(arr,N);
-    printf("Executed in %lf ms\n",time);
-
-    cudaFree(d_arr);
-    free(arr);
-    return 0;
-}
+void print(int arr[], int n) {
+    for (long i = 0; i < n; i++) {
+      printf("%d \n",arr[i]);
 
+   }
+   printf("\n");
+}
\ No newline at end of file