llvm/offload/test/offloading/bug64959.c

// RUN: %libomptarget-compile-run-and-check-generic
// RUN: %libomptarget-compileopt-run-and-check-generic

// TODO: This requires malloc support for the threads states.
// FIXME: Flaky on all GPU targets.
// UNSUPPORTED: amdgcn-amd-amdhsa
// UNSUPPORTED: nvptx64-nvidia-cuda
// UNSUPPORTED: nvptx64-nvidia-cuda-LTO

#include <omp.h>
#include <stdio.h>
#define N 10

int isCPU() { return 1; }

#pragma omp begin declare variant match(device = {kind(gpu)})
int isCPU() { return 0; }
#pragma omp end declare variant

int main(void) {
  long int aa = 0;
  int res = 0;

  int ng = 12;
  int cmom = 14;
  int nxyz;

#pragma omp target map(from : nxyz, ng, cmom)
  {
    nxyz = isCPU() ? 2 : 5000;
    ng = isCPU() ? 2 : 12;
    cmom = isCPU() ? 2 : 14;
  }

#pragma omp target teams distribute num_teams(nxyz)                            \
    thread_limit(ng *(cmom - 1)) map(tofrom : aa)
  for (int gid = 0; gid < nxyz; gid++) {
#pragma omp parallel for collapse(2)
    for (unsigned int g = 0; g < ng; g++) {
      for (unsigned int l = 0; l < cmom - 1; l++) {
        int a = 0;
#pragma omp parallel for reduction(+ : a)
        for (int i = 0; i < N; i++) {
          a += i;
        }
#pragma omp atomic
        aa += a;
      }
    }
  }
  long exp = (long)ng * (cmom - 1) * nxyz * (N * (N - 1) / 2);
  printf("The result is = %ld exp:%ld!\n", aa, exp);
  if (aa != exp) {
    printf("Failed %ld\n", aa);
    return 1;
  }
  // CHECK: Success
  printf("Success\n");
  return 0;
}