SSE vs Native Loop over FLOAT

main.c

#include <xmmintrin.h>
#include <stdlib.h>
#include <stdio.h>
#include <time.h>

void __attribute__ ((noinline)) SSE(){
	int len = 1024;
	float data1 [1024] = {1, 2, 3, 4, 5, 6, 7, 8};
	float data2 [1024] = {1, 2, 3, 4, 5, 6, 7, 8};
	float dest[1024];
	
	unsigned int i = 0;
	
	for(; i < len; i+=4){
		__m128 sse1 = _mm_load_ps(data1 +i);
		__m128 sse2 = _mm_load_ps(data2 + i);
		__m128 sseRes = _mm_mul_ps(sse1, sse2);
		_mm_store_ps(dest + i, sseRes);
	}
}

void __attribute__ ((noinline)) std(){
	int len = 1024;
	unsigned int i = 0;
	float data1 [1024] = {1, 2, 3, 4, 5, 6, 7, 8};
	float data2 [1024] = {1, 2, 3, 4, 5, 6, 7, 8};
	float dest[1024] = {0};
	
	for(; i<len; i++){
		dest[i] = data1[i] * data2[i];
	}
}

int main(int argc, char *argv[]) {
	printf("Hello, world!\n");
	int i = 0;
	clock_t begin = clock();
	for(;i<1000000;i++){
		SSE();
	}
	clock_t end = clock();
	double time_spent = (double)(end - begin) / CLOCKS_PER_SEC;
	printf("SSE: %f\n", time_spent);
	
	clock_t begin2= clock();
	int j = 0;
	for(j=0;j<1000000;j++){
		std();
	}
	clock_t end2 = clock();
	double time_spent2 = (double)(end2 - begin2) / CLOCKS_PER_SEC;
	printf("std: %f\n", time_spent2);
	printf("bye!\n");
	
	return 0;
}