[OpenMP] Host offload reduction in parallel regions hangs

[jdoerfert]

A host offload reduction in an outer parallel regions seems to hang.
Here is the test:

#include <iostream>
#include <vector>

#define N 8

int main() {
  std::vector<int> avec(N);
  int *a = avec.data();
#pragma omp parallel for
  for (int i = 0; i < N; i++) {
    a[i] = 0;
#pragma omp target teams distribute parallel for reduction(+ : a[i])
    for (int j = 0; j < N; j++)
      a[i] += 1;
  }

  // CHECK: 8
  // CHECK: 8
  // CHECK: 8
  // CHECK: 8
  // CHECK: 8
  // CHECK: 8
  // CHECK: 8
  // CHECK: 8
  for (int i = 0; i < N; i++)
    std::cout << a[i] << std::endl;
}

Even with N=1 it hangs.

[llvmbot]

@llvm/issue-subscribers-openmp

Author: Johannes Doerfert (jdoerfert)

A host offload reduction in an outer parallel regions seems to hang. Here is the test: ``` #include <iostream> #include <vector>

#define N 8

int main() {
std::vector<int> avec(N);
int *a = avec.data();
#pragma omp parallel for
for (int i = 0; i < N; i++) {
a[i] = 0;
#pragma omp target teams distribute parallel for reduction(+ : a[i])
for (int j = 0; j < N; j++)
a[i] += 1;
}

// CHECK: 8
// CHECK: 8
// CHECK: 8
// CHECK: 8
// CHECK: 8
// CHECK: 8
// CHECK: 8
// CHECK: 8
for (int i = 0; i < N; i++)
std::cout << a[i] << std::endl;
}

Even with N=1 it hangs.
</details>