1  /* { dg-do run { target openacc_nvidia_accel_selected } } */
       2  /* { dg-additional-options "-foffload=-fdump-tree-oaccloops" } */
       3  /* { dg-set-target-env-var "GOMP_DEBUG" "1" } */
       4  
       5  #include <stdlib.h>
       6  
       7  #define N 1024
       8  
       9  unsigned int a[N];
      10  unsigned int b[N];
      11  unsigned int c[N];
      12  unsigned int n = N;
      13  
      14  int
      15  main (void)
      16  {
      17    for (unsigned int i = 0; i < n; ++i)
      18      {
      19        a[i] = i % 3;
      20        b[i] = i % 5;
      21      }
      22  
      23  #pragma acc parallel vector_length (128) copyin (a,b) copyout (c)
      24    {
      25  #pragma acc loop worker
      26      for (unsigned int i = 0; i < 4; i++)
      27  #pragma acc loop vector
      28        for (unsigned int j = 0; j < n / 4; j++)
      29  	c[(i * N / 4) + j] = a[(i * N / 4) + j] + b[(i * N / 4) + j];
      30    }
      31  
      32    for (unsigned int i = 0; i < n; ++i)
      33      if (c[i] != (i % 3) + (i % 5))
      34        abort ();
      35  
      36    return 0;
      37  }
      38  
      39  /* { dg-final { scan-offload-tree-dump "__attribute__\\(\\(oacc function \\(1, 0, 128\\)" "oaccloops" } } */
      40  /* { dg-output "nvptx_exec: kernel main\\\$_omp_fn\\\$0: launch gangs=1, workers=\[1-8\], vectors=128" } */