Search⌘ K
AI Features
Home
Courses
Master Modern C++

Solution: Safe vs. Unsafe Parallelism

Understand how std::reduce combined with std::execution::par enables safe parallel computations in C++. Discover why std::reduce avoids data races by working on partial sums and merging results thread-safely. This lesson helps you grasp safe parallel algorithm implementation for improved performance without risking unsafe shared state modifications.

We'll cover the following...
C++ 23
#include <iostream>
#include <vector>
#include <algorithm>
#include <execution>
#include <numeric>
int main() {
    std::vector<int> sensor_readings(10'000'000, 1);
    // We replace the unsafe side-effect loop with std::reduce
    // std::reduce automatically handles partitioning and combining results 
    // from multiple threads safely
    long long total_energy = std::reduce(
        std::execution::par, 
        sensor_readings.begin(), 
        sensor_readings.end()
    );
    std::cout << "Total Energy: " << total_energy << "\n";
    std::cout << "Expected:     10000000\n";
    if (total_energy == 10'000'000) {
        std::cout << "Status: Success (Safe Parallelism)\n";
    } else {
        std::cout << "Status: Failure (Incorrect Result)\n";
    }
    return 0;
}
...