void sgemm(int m, int n, float *A, float *C) {

	__m128 r1, r2, r3, r4, r5, r6, r7, r8, r9, r10, r11, r12, r13, r14, r15, r16;

	__m128 temp;

	//Load column from original
	for (int i = 0; i < n; i++) {
		r1 = _mm_loadu_ps(A+i*n+0);
		r2 = _mm_loadu_ps(A+i*n+4);
		r3 = _mm_loadu_ps(A+i*n+8);
		r4 = _mm_loadu_ps(A+i*n+12);
		r5 = _mm_loadu_ps(A+i*n+16);
		r6 = _mm_loadu_ps(A+i*n+20);
		r7 = _mm_loadu_ps(A+i*n+24);
		r8 = _mm_loadu_ps(A+i*n+28);
		r9 = _mm_loadu_ps(A+i*n+32);
		r10 = _mm_loadu_ps(A+i*n+36);
		r11 = _mm_loadu_ps(A+i*n+40);
		r12 = _mm_loadu_ps(A+i*n+44);
		r13 = _mm_loadu_ps(A+i*n+48);
		r14 = _mm_loadu_ps(A+i*n+52);
		r15 = _mm_loadu_ps(A+i*n+56);
		r16 = _mm_loadu_ps(A+i*n+60);
		for (int j = 0; j < 64; j++) {
			temp = _mm_load1_ps(A+j+i*n); //can improve this further, (float *)&r1 + j
			_mm_storeu_ps(C+j*n+0, _mm_add_ps(_mm_loadu_ps(C+j*n+0), _mm_mul_ps(r1, temp)));
			_mm_storeu_ps(C+j*n+4, _mm_add_ps(_mm_loadu_ps(C+j*n+4), _mm_mul_ps(r2, temp)));
			_mm_storeu_ps(C+j*n+8, _mm_add_ps(_mm_loadu_ps(C+j*n+8), _mm_mul_ps(r3, temp)));
			_mm_storeu_ps(C+j*n+12, _mm_add_ps(_mm_loadu_ps(C+j*n+12), _mm_mul_ps(r4, temp)));
			_mm_storeu_ps(C+j*n+16, _mm_add_ps(_mm_loadu_ps(C+j*n+16), _mm_mul_ps(r5, temp)));
			_mm_storeu_ps(C+j*n+20, _mm_add_ps(_mm_loadu_ps(C+j*n+20), _mm_mul_ps(r6, temp)));
			_mm_storeu_ps(C+j*n+24, _mm_add_ps(_mm_loadu_ps(C+j*n+24), _mm_mul_ps(r7, temp)));
			_mm_storeu_ps(C+j*n+28, _mm_add_ps(_mm_loadu_ps(C+j*n+28), _mm_mul_ps(r8, temp)));
			_mm_storeu_ps(C+j*n+32, _mm_add_ps(_mm_loadu_ps(C+j*n+32), _mm_mul_ps(r9, temp)));
			_mm_storeu_ps(C+j*n+36, _mm_add_ps(_mm_loadu_ps(C+j*n+36), _mm_mul_ps(r10, temp)));
			_mm_storeu_ps(C+j*n+40, _mm_add_ps(_mm_loadu_ps(C+j*n+40), _mm_mul_ps(r11, temp)));
			_mm_storeu_ps(C+j*n+44, _mm_add_ps(_mm_loadu_ps(C+j*n+44), _mm_mul_ps(r12, temp)));
			_mm_storeu_ps(C+j*n+48, _mm_add_ps(_mm_loadu_ps(C+j*n+48), _mm_mul_ps(r13, temp)));
			_mm_storeu_ps(C+j*n+52, _mm_add_ps(_mm_loadu_ps(C+j*n+52), _mm_mul_ps(r14, temp)));
			_mm_storeu_ps(C+j*n+56, _mm_add_ps(_mm_loadu_ps(C+j*n+56), _mm_mul_ps(r15, temp)));
			_mm_storeu_ps(C+j*n+60, _mm_add_ps(_mm_loadu_ps(C+j*n+60), _mm_mul_ps(r16, temp)));
		}
	}

}