#include <stdio.h>
#include <ctime>
#include <arm_neon.h> //启用 NEON 指令
#include <stdlib.h>

#define SIZE 1024

void transposeMatrix(float** matrix, float** transposed, int size) {
    for (int i = 0; i < size; i++) {
        for (int j = 0; j < size; j++) {
            transposed[j][i] = matrix[i][j];
        }
    }
}

void matmul(float** A, float** B, float** C, int n) {
    for (int i = 0; i < n; i++) {
        for (int j = 0; j < n; j++) {
            float sum = 0.0f;
            for (int k = 0; k < n; k += 4) 
            {
                float32x4_t vecA, vecB, vecC;
                if (k + 4 <= n) 
                {
                    // 加载A和B的4个元素，进行向量化计算
                    vecA = vld1q_f32(&A[i][k]);
                    vecB = vld1q_f32(&B[j][k]);
                    // 向量化乘法并累加结果
                    vecC = vmulq_f32(vecA, vecB);
                    sum += vgetq_lane_f32(vecC, 0) + vgetq_lane_f32(vecC, 1) +
                           vgetq_lane_f32(vecC, 2) + vgetq_lane_f32(vecC, 3);
                } 
                else 
                {
                    // 处理剩余的元素
                    for (int m = k; m < n; m++) 
                    {
                        sum += A[i][m] * B[j][m];
                    }
                }
            }
            C[i][j] = sum;
        }
    }
}

int main() {
    srand(time(NULL));

    float** A = (float**)malloc(SIZE * sizeof(float*));
    float** B = (float**)malloc(SIZE * sizeof(float*));
    float** C = (float**)malloc(SIZE * sizeof(float*));
    float** transposed = (float**)malloc(SIZE * sizeof(float*));

    for (int i = 0; i < SIZE; i++) {
        A[i] = (float*)malloc(SIZE * sizeof(float));
        B[i] = (float*)malloc(SIZE * sizeof(float));
        C[i] = (float*)malloc(SIZE * sizeof(float));
        transposed[i] = (float*)malloc(SIZE * sizeof(float));
    }

    for (int i = 0; i < SIZE; i++) {
        for (int j = 0; j < SIZE; j++) {
            A[i][j] = (float)(rand() % 100) / 100.0f;
            B[i][j] = (float)(rand() % 100) / 100.0f;
        }
    }

    transposeMatrix(B, transposed, SIZE);

    clock_t start = clock();
    matmul(A, transposed, C, SIZE);
    clock_t end = clock();

    // 计算并输出矩阵乘法的时间
    double multiply_time_spent = double(end - start) / CLOCKS_PER_SEC;
    printf("使用优化的向量乘法：\n当SIZE取%d时，优化的向量乘法时间：%lf秒\n", SIZE, multiply_time_spent);


    for (int i = 0; i < SIZE; i++) {
        free(A[i]);
        free(B[i]);
        free(C[i]);
        free(transposed[i]);
    }
    free(A);
    free(B);
    free(C);
    free(transposed);

    return 0;
}