#include <arm_neon.h>
#include <stdio.h>
#include <time.h>
#include <stdlib.h>
#define SIZE 1024

void vector_add_optimized(float* A, float* B, float* C, int size);
int main() {
	float A[SIZE], B[SIZE], C[SIZE];
	int i;
	srand(time(0));
	for(i=0; i<SIZE; i++) {
		A[i] = rand() % 100;
		B[i] = rand() % 100;
	}

	clock_t start = clock();
	vector_add_optimized(A, B, C, SIZE);
	clock_t end = clock();
	printf("使用NEON 优化向量加法时间：%lf\n", (double)(end-start) / CLOCKS_PER_SEC);
}

void vector_add_optimized(float* A, float* B, float* C, int size) {
	int i;
	for(i=0; i<size; i+=4) {
		float32x4_t vecA = vld1q_f32(&A[i]);
		float32x4_t vecB = vld1q_f32(&B[i]);
		float32x4_t vecC = vaddq_f32(vecA, vecB);
		vst1q_f32(&C[i], vecC);
	}
}