youmuliu/stgsharp
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

daily update on 20230829 by nitload youmu

Browse Source
This commit is contained in:
Nitload
2023-08-29 22:29:47 +08:00
parent 8aac9fee87
commit 1fcbfb972b
125 changed files with 6505 additions and 2077 deletions
Download Patch File Download Diff File Expand all files Collapse all files
源代码/StgShqrpGraphic/mat3.c
+39 -11
View File
@@ -1,14 +1,47 @@
#include "StgSharpGraphic.h"
#include <stdio.h>
SSGC_API __inline void _cdecl transpose3to2(matrix3map* source, matrix2map* target)
{
__m128 t0 = _mm_shuffle_ps(ALIGN(source->colum[0]), ALIGN(source->colum[1]), _MM_SHUFFLE(1, 0, 1, 0));
__m128 t2 = _mm_shuffle_ps(ALIGN(source->colum[2]), *defaultVecPtr, _MM_SHUFFLE(1, 0, 1, 0));
target->colum[0] = _mm_shuffle_ps(t0, t2, _MM_SHUFFLE(2, 0, 2, 0));
target->colum[1] = _mm_shuffle_ps(t0, t2, _MM_SHUFFLE(3, 1, 3, 1));
}
SSGC_API __inline void _cdecl transpose3to3(matrix3map* source, matrix3map* target)
{
__m128 t0 = _mm_shuffle_ps(ALIGN(source->colum[0]), ALIGN(source->colum[1]), _MM_SHUFFLE(1, 0, 1, 0));
__m128 t1 = _mm_shuffle_ps(ALIGN(source->colum[0]), ALIGN(source->colum[1]), _MM_SHUFFLE(3, 2, 3, 2));
__m128 t2 = _mm_shuffle_ps(ALIGN(source->colum[2]), *defaultVecPtr, _MM_SHUFFLE(1, 0, 1, 0));
__m128 t3 = _mm_shuffle_ps(ALIGN(source->colum[2]), *defaultVecPtr, _MM_SHUFFLE(3, 2, 3, 2));
target->colum[0] = _mm_shuffle_ps(t0, t2, _MM_SHUFFLE(2, 0, 2, 0));
target->colum[1] = _mm_shuffle_ps(t0, t2, _MM_SHUFFLE(3, 1, 3, 1));
target->colum[2] = _mm_shuffle_ps(t1, t3, _MM_SHUFFLE(2, 0, 2, 0));
}
SSGC_API __inline void _cdecl transpose3to4(matrix3map* source, matrix4map* target)
{
__m128 t0 = _mm_shuffle_ps(ALIGN(source->colum[0]), ALIGN(source->colum[1]), _MM_SHUFFLE(1, 0, 1, 0));
__m128 t1 = _mm_shuffle_ps(ALIGN(source->colum[0]), ALIGN(source->colum[1]), _MM_SHUFFLE(3, 2, 3, 2));
__m128 t2 = _mm_shuffle_ps(ALIGN(source->colum[2]), *defaultVecPtr, _MM_SHUFFLE(1, 0, 1, 0));
__m128 t3 = _mm_shuffle_ps(ALIGN(source->colum[2]), *defaultVecPtr, _MM_SHUFFLE(3, 2, 3, 2));
target->colum[0] = _mm_shuffle_ps(t0, t2, _MM_SHUFFLE(2, 0, 2, 0));
target->colum[1] = _mm_shuffle_ps(t0, t2, _MM_SHUFFLE(3, 1, 3, 1));
target->colum[2] = _mm_shuffle_ps(t1, t3, _MM_SHUFFLE(2, 0, 2, 0));
target->colum[3] = _mm_shuffle_ps(t1, t3, _MM_SHUFFLE(3, 1, 3, 1));
}
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]);
__m128 addc2 = _mm_shuffle_ps(c2, c2, _MM_SHUFFLE(3, 1, 0, 2));
//let sub row aligned
__m128 subc1 = _mm_shuffle_ps(c1, c1, _MM_SHUFFLE(3, 1, 0, 2));
@@ -18,13 +51,9 @@ float det_mat3_internal(__m128 const c0, __m128 const c1, __m128 const c2)
__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]);
sub = _mm_mul_ps(sub, subc2);
//main-corner sub sub-corner
cache = _mm_sub_ps(cache,sub);
@@ -32,7 +61,6 @@ float det_mat3_internal(__m128 const c0, __m128 const c1, __m128 const c2)
__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;
return ret.m128_f32[2];
}