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

daily update on 20241002 by nitload youmu, intrinsic function rebuild, syntax style arrange

Browse Source
This commit is contained in:
Nitload
2024-10-03 04:03:06 +08:00
parent 0979123ac0
commit 6e2f93b053
265 changed files with 26732 additions and 19058 deletions
Download Patch File Download Diff File Expand all files Collapse all files
源代码/StgSharpC/mat4.c
+57 -43
View File
@@ -1,62 +1,76 @@
#include "StgSharpC.h"
#include "ssgc_internal.h"
#include "ssc_intrinsic.h"
#include <stdio.h>
#include <pmmintrin.h>
SSCAPI __inline void __cdecl transpose4to2(matrix4map* source, matrix4map* target)
INTERNAL void __cdecl transpose42(__4_columnset *source, __2_columnset *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]), ALIGN(source->colum[3]), _MM_SHUFFLE(1, 0, 1, 0));
__m128 t0 = _mm_shuffle_ps(ALIGN(source->column[0]),
ALIGN(source->column[1]),
_MM_SHUFFLE(1, 0, 1, 0));
__m128 t2 = _mm_shuffle_ps(ALIGN(source->column[2]),
ALIGN(source->column[3]),
_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));
target->column[0] = _mm_shuffle_ps(t0, t2, _MM_SHUFFLE(2, 0, 2, 0));
target->column[1] = _mm_shuffle_ps(t0, t2, _MM_SHUFFLE(3, 1, 3, 1));
}
SSCAPI __inline void __cdecl transpose4to4(matrix4map* source, matrix4map* target)
INTERNAL void __cdecl transpose43(__4_columnset *source, __3_columnset *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]), ALIGN(source->colum[3]), _MM_SHUFFLE(1, 0, 1, 0));
__m128 t3 = _mm_shuffle_ps(ALIGN(source->colum[2]), ALIGN(source->colum[3]), _MM_SHUFFLE(3, 2, 3, 2));
__m128 t0 = _mm_shuffle_ps(ALIGN(source->column[0]),
ALIGN(source->column[1]),
_MM_SHUFFLE(1, 0, 1, 0));
__m128 t1 = _mm_shuffle_ps(ALIGN(source->column[0]),
ALIGN(source->column[1]),
_MM_SHUFFLE(3, 2, 3, 2));
__m128 t2 = _mm_shuffle_ps(ALIGN(source->column[2]),
ALIGN(source->column[3]),
_MM_SHUFFLE(1, 0, 1, 0));
__m128 t3 = _mm_shuffle_ps(ALIGN(source->column[2]),
ALIGN(source->column[3]),
_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));
target->column[0] = _mm_shuffle_ps(t0, t2, _MM_SHUFFLE(2, 0, 2, 0));
target->column[1] = _mm_shuffle_ps(t0, t2, _MM_SHUFFLE(3, 1, 3, 1));
target->column[2] = _mm_shuffle_ps(t1, t3, _MM_SHUFFLE(2, 0, 2, 0));
}
SSCAPI __inline void __cdecl transpose4to3(matrix4map* source, matrix3map* target)
INTERNAL void __cdecl transpose44(__4_columnset *source, __4_columnset *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]), ALIGN(source->colum[3]), _MM_SHUFFLE(1, 0, 1, 0));
__m128 t3 = _mm_shuffle_ps(ALIGN(source->colum[2]), ALIGN(source->colum[3]), _MM_SHUFFLE(3, 2, 3, 2));
__m128 t0 = _mm_shuffle_ps(ALIGN(source->column[0]),
ALIGN(source->column[1]),
_MM_SHUFFLE(1, 0, 1, 0));
__m128 t1 = _mm_shuffle_ps(ALIGN(source->column[0]),
ALIGN(source->column[1]),
_MM_SHUFFLE(3, 2, 3, 2));
__m128 t2 = _mm_shuffle_ps(ALIGN(source->column[2]),
ALIGN(source->column[3]),
_MM_SHUFFLE(1, 0, 1, 0));
__m128 t3 = _mm_shuffle_ps(ALIGN(source->column[2]),
ALIGN(source->column[3]),
_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->column[0] = _mm_shuffle_ps(t0, t2, _MM_SHUFFLE(2, 0, 2, 0));
target->column[1] = _mm_shuffle_ps(t0, t2, _MM_SHUFFLE(3, 1, 3, 1));
target->column[2] = _mm_shuffle_ps(t1, t3, _MM_SHUFFLE(2, 0, 2, 0));
target->column[3] = _mm_shuffle_ps(t1, t3, _MM_SHUFFLE(3, 1, 3, 1));
}
SSCAPI __inline float __cdecl det_mat4(matrix4map* matPtr, matrix4map* transpose)
INTERNAL float __cdecl det_mat4(__4_columnset *matPtr, __4_columnset *transpose)
{
__m128 mc0 = ALIGN(matPtr->colum[0]);
__m128 mc1 = ALIGN(matPtr->colum[1]);
__m128 mc2 = ALIGN(matPtr->colum[2]);
__m128 mc3 = ALIGN(matPtr->colum[3]);
__m128 mc0 = ALIGN(matPtr->column[0]);
__m128 mc1 = ALIGN(matPtr->column[1]);
__m128 mc2 = ALIGN(matPtr->column[2]);
__m128 mc3 = ALIGN(matPtr->column[3]);
__m128 cogeneration = _mm_set_ps(
det_mat3_internal(mc0, mc1, mc2),
det_mat3_internal(mc0, mc1, mc3),
det_mat3_internal(mc0, mc2, mc3),
det_mat3_internal(mc1, mc2, mc3)
);
__m128 retvec = _mm_mul_ps(ALIGN(transpose->colum[3]), cogeneration);
retvec = _mm_hsub_ps(retvec, retvec);
retvec = _mm_hadd_ps(retvec, retvec);
return -retvec.m128_f32[0];
__m128 cogeneration = _mm_set_ps(det_mat3_internal(mc0, mc1, mc2),
det_mat3_internal(mc0, mc1, mc3),
det_mat3_internal(mc0, mc2, mc3),
det_mat3_internal(mc1, mc2, mc3));
__m128 retvec = _mm_mul_ps(ALIGN(transpose->column[3]), cogeneration);
retvec = _mm_hsub_ps(retvec, retvec);
retvec = _mm_hadd_ps(retvec, retvec);
return -retvec.m128_f32[0];
}
SSCAPI __inline void __cdecl deinit_mat4(matrix4map* matPtr)
{
_aligned_free(matPtr);
}