39 lines
1.3 KiB
C
39 lines
1.3 KiB
C
#include "StgSharpGraphic.h"
|
|
#include <stdio.h>
|
|
|
|
float det_mat3_internal(__m128 const c0, __m128 const c1, __m128 const c2)
|
|
{
|
|
//let adding row aligned
|
|
__m128 addc1 = _mm_shuffle_ps(c1, c1, _MM_SHUFFLE(3, 0, 2, 1));
|
|
__m128 addc2 = _mm_shuffle_ps(c2, c1, _MM_SHUFFLE(3, 1, 0, 2));
|
|
|
|
printf("%f\t%f\t%f\n", addc1.m128_f32[0], addc1.m128_f32[1], addc1.m128_f32[2]);
|
|
printf("%f\t%f\t%f\n\n", addc2.m128_f32[0], addc2.m128_f32[1], addc2.m128_f32[2]);
|
|
|
|
//let sub row aligned
|
|
__m128 subc1 = _mm_shuffle_ps(c1, c1, _MM_SHUFFLE(3, 1, 0, 2));
|
|
__m128 subc2 = _mm_shuffle_ps(c2, c2, _MM_SHUFFLE(3, 0, 2, 1));
|
|
|
|
//multple all add row
|
|
__m128 cache = _mm_mul_ps(c0,addc1);
|
|
cache = _mm_mul_ps(cache,addc2);
|
|
|
|
printf("%f\t%f\t%f\n\n", cache.m128_f32[0], cache.m128_f32[1], cache.m128_f32[2]);
|
|
|
|
//multiple all sub row
|
|
__m128 sub = _mm_mul_ps(c0, subc1);
|
|
sub = _mm_mul_ps(sub, subc1);
|
|
|
|
printf("%f\t%f\t%f\n\n", sub.m128_f32[0], sub.m128_f32[1], sub.m128_f32[2]);
|
|
|
|
//main-corner sub sub-corner
|
|
cache = _mm_sub_ps(cache,sub);
|
|
|
|
__m128 ret = _mm_hadd_ps(cache,cache); //element 1 add element 2
|
|
ret = _mm_add_ps(ret, cache); //(e1 add e2) add element 3
|
|
|
|
int a = *(int*)0; //for debug, all an excepotion
|
|
|
|
return ret.m128_f32[3] + a;
|
|
}
|