#include <stdio.h>
#include <stdlib.h>
#include <ctime>
#include <arm_neon.h>
// 定义向量大小
#define SIZE 1024
// 原始的向量加法函数
void vector_add(float* A, float* B, float* C, int size) {
    for (int i = 0; i < size; i++) {
        C[i] = A[i] + B[i];
    }
}
// 使用NEON指令优化的向量加法函数
void vector_add_optimized(float* A, float* B, float* C, int size) {
    int i;
    for (i = 0; i < size - 3; i += 4) {
        // 向量加载，将A和B的4个连续元素加载到float32x4_t类型的向量中
        float32x4_t a_vec = vld1q_f32(&A[i]);
        float32x4_t b_vec = vld1q_f32(&B[i]);
        float32x4_t c_vec = vaddq_f32(a_vec, b_vec);
        // 将结果存储到C中
        vst1q_f32(&C[i], c_vec);
    }
    for (; i < size; i++) {
        C[i] = A[i] + B[i];
    }
}
int main() {
    float A[SIZE];
    float B[SIZE];
    float C[SIZE];
    float C_optimized[SIZE];

    // 利用for循环将A和B向量的每个元素随机初始化
    for (int i = 0; i < SIZE; i++) {
        A[i] = (float)(rand() % 100);
        B[i] = (float)(rand() % 100);
    }

    // 测试原始向量加法函数的运行时间
    clock_t start_time_original = clock();
    vector_add(A, B, C, SIZE);
    clock_t end_time_original = clock();
    double elapsed_time_original = (double)(end_time_original - start_time_original) / CLOCKS_PER_SEC;

    clock_t start_time_optimized = clock();
    vector_add_optimized(A, B, C_optimized, SIZE);
    clock_t end_time_optimized = clock();
    double elapsed_time_optimized = (double)(end_time_optimized - start_time_optimized) / CLOCKS_PER_SEC;
     
    printf("original time: %lf s\n", elapsed_time_original);
    // 输出NEON优化后的向量加法的运行时间
    printf("NEON optimized time: %lf s\n", elapsed_time_optimized);
    return 0;
}