How to write parallel computing in C programming language

1 year ago
Sophia Anderson
2 minutes

In C programming language, parallel computing can be achieved using multithreading. Below is a simple example code demonstrating how to perform parallel computing using multithreading.

#include <stdio.h>
#include <pthread.h>

#define THREADS_COUNT 4
#define ARRAY_SIZE 1000000

int array[ARRAY_SIZE];
int sum = 0;

// 线程函数，用于计算数组的部分和
void* calculateSum(void* arg) {
    int thread_id = *(int*)arg;
    int start = thread_id * (ARRAY_SIZE / THREADS_COUNT);
    int end = start + (ARRAY_SIZE / THREADS_COUNT);

    for (int i = start; i < end; i++) {
        sum += array[i];
    }

    return NULL;
}

int main() {
    // 初始化数组
    for (int i = 0; i < ARRAY_SIZE; i++) {
        array[i] = i;
    }

    pthread_t threads[THREADS_COUNT];
    int thread_ids[THREADS_COUNT];

    // 创建多个线程，每个线程负责计算数组的一部分
    for (int i = 0; i < THREADS_COUNT; i++) {
        thread_ids[i] = i;
        pthread_create(&threads[i], NULL, calculateSum, &thread_ids[i]);
    }

    // 等待所有线程执行完毕
    for (int i = 0; i < THREADS_COUNT; i++) {
        pthread_join(threads[i], NULL);
    }

    printf("Sum: %d\n", sum);

    return 0;
}

In the above code, we define an integer array ‘array’ containing one million elements. Then, we create four threads, with each thread calculating the sum of a portion of the array. Finally, we add up the calculated partial sums to get the final result and output it.

In parallel computing, we use mutex locks to ensure thread safety by preventing multiple threads from simultaneously modifying the sum variable to avoid race conditions. In this simple example, using mutex locks is not necessary because each thread calculates a non-overlapping part, and they do not access the same memory location at the same time. However, in more complex parallel computing scenarios, mutex locks may be needed to ensure data consistency.

Additionally, it is important to note that using multiple threads may not necessarily improve a program’s performance, as the overhead of switching and syncing between threads may outweigh the performance gains from parallel computing. Therefore, when using multiple threads for parallel computing, it is important to evaluate and optimize based on the specific application scenario.

bannerAds