int main() {
#pragma omp parallel for
for (int i = 0; i < 1000; i = i + 1) {
std::cout << "some";
}
//other code
}
Valgrind report:
possibly lost: 2,128 bytes in 7 blocks
If I omit the "other code" in the example above everything works fine. Unfortunately I can not share the code. Is there maybe a common pattern for why this can happen? Note that in the "other code" no parallelism is involved.
Nb: I also tried something else in the loop than cout and it was the same issue.
EDIT: Maybe it goes without saying but to make sure: Dropping #pragma omp parallel for results in a report without any leaks.
