Question 14
Back to Exercise
What is the output of the following C++ program? (The program is free of data races, so its output is the same on every run.)
#include <barrier>
#include <iostream>
#include <thread>
#include <vector>
int main() {
const int n = 3;
std::vector<int> a(n), b(n);
int phases = 0;
std::barrier sync(n, [&]() noexcept { ++phases; }); // completion step: runs once per phase
std::vector<std::jthread> pool;
for (int i = 0; i < n; ++i)
pool.emplace_back([&, i] {
a[i] = i + 1; // phase 1
sync.arrive_and_wait();
b[i] = a[0] + a[1] + a[2]; // phase 2 reads everyone's phase-1 result
sync.arrive_and_wait();
});
pool.clear();
std::cout << b[0] << b[1] << b[2] << ' ' << phases << '\n';
return 0;
}
- A 666 2 Correct
- B 123 2
- C 666 1
- D 666 6
Correct Answer: Option A
No thread starts phase 2 until all have finished phase 1, so every b[i] = 1 + 2 + 3. The completion function runs once at the end of each of the two phases.
Discussion & Comments 0
Share your method, a shortcut, or ask about a step you didn't follow.
No comments yet. Be the first to share your solution!