External/cglm
1
0
Fork 0
mirror of https://github.com/recp/cglm.git synced 2025-10-04 09:08:53 +00:00
Code Issues Packages Projects Releases Wiki Activity

drop manual simd impl from mat3

Browse Source
This commit is contained in:
Recep Aslantas
2016-10-11 18:26:07 +03:00
parent 3c26a7f820
commit 80c6c0a28b
4 changed files with 0 additions and 95 deletions
Download Patch File Download Diff File Expand all files Collapse all files

19
include/cglm-mat-simd-sse2.h
View File

@@ -11,25 +11,6 @@
#include "cglm-intrin.h"
CGLM_INLINE
void
glm_mat4_pick3_sse2(mat4 mat, mat3 dest) {
__m128 x0, x1, x2, x3;
x0 = _mm_load_ps(mat[0]);
x1 = _mm_load_ps(mat[1]);
x2 = _mm_load_ps(mat[2]);
x3 = _mm_shuffle_ps(x0, x1, _MM_SHUFFLE(0, 0, 2, 2));
x0 = _mm_shuffle_ps(x0, x3, _MM_SHUFFLE(2, 0, 1, 0));
x1 = _mm_shuffle_ps(x1, x2, _MM_SHUFFLE(1, 0, 2, 1));
_mm_storeu_ps(&dest[0][0], x0);
_mm_storeu_ps(&dest[1][1], x1);
dest[2][2] = mat[2][2];
}
CGLM_INLINE
void
glm_mat4_scale_sse2(mat4 m, float s){

4
include/cglm-mat.h
View File

@@ -72,9 +72,6 @@ glm_mat4_dup(mat4 mat, mat4 dest) {
CGLM_INLINE
void
glm_mat4_pick3(mat4 mat, mat3 dest) {
#if defined( __SSE__ ) || defined( __SSE2__ )
glm_mat4_pick3_sse2(mat, dest);
#else
dest[0][0] = mat[0][0];
dest[0][1] = mat[0][1];
dest[0][2] = mat[0][2];
@@ -86,7 +83,6 @@ glm_mat4_pick3(mat4 mat, mat3 dest) {
dest[2][0] = mat[2][0];
dest[2][1] = mat[2][1];
dest[2][2] = mat[2][2];
#endif
}
/*!

60
include/cglm-mat3-simd-sse2.h
View File

@@ -11,66 +11,6 @@
#include "cglm-intrin.h"
CGLM_INLINE
void
glm_mat3_transp_to_sse2(mat3 m, mat3 dest){
__m128 x0, x1, x2, x3, x4;
/*
a b c d a d g b
e f g h -> e h c f
j j
*/
/* d c b a */
/* h g f e */
x0 = _mm_loadu_ps(&m[0][0]);
x1 = _mm_loadu_ps(&m[1][1]);
/* g g b b */
/* a d g b */
x2 = _mm_shuffle_ps(x0, x1, _MM_SHUFFLE(2, 2, 1, 1));
x3 = _mm_shuffle_ps(x0, x2, _MM_SHUFFLE(0, 2, 3, 0));
/* c c f f */
/* e h c f */
x2 = _mm_shuffle_ps(x1, x0, _MM_SHUFFLE(2, 2, 1, 1));
x4 = _mm_shuffle_ps(x1, x2, _MM_SHUFFLE(0, 2, 3, 0));
_mm_storeu_ps(&dest[0][0], x3);
_mm_storeu_ps(&dest[1][1], x4);
dest[2][2] = m[2][2];
}
CGLM_INLINE
void
glm_mat3_transp_sse2(mat3 m){
__m128 x0, x1, x2, x3, x4;
/*
a b c d a d g b
e f g h -> e h c f
j j
*/
/* d c b a */
/* h g f e */
x0 = _mm_loadu_ps(&m[0][0]);
x1 = _mm_loadu_ps(&m[1][1]);
/* g g b b */
/* a d g b */
x2 = _mm_shuffle_ps(x0, x1, _MM_SHUFFLE(2, 2, 1, 1));
x3 = _mm_shuffle_ps(x0, x2, _MM_SHUFFLE(0, 2, 3, 0));
/* c c f f */
/* e h c f */
x2 = _mm_shuffle_ps(x1, x0, _MM_SHUFFLE(2, 2, 1, 1));
x4 = _mm_shuffle_ps(x1, x2, _MM_SHUFFLE(0, 2, 3, 0));
_mm_storeu_ps(&m[0][0], x3);
_mm_storeu_ps(&m[1][1], x4);
}
CGLM_INLINE
void
glm_mat3_mul_sse2(mat3 m1, mat3 m2, mat3 dest) {

12
include/cglm-mat3.h
View File

@@ -66,12 +66,10 @@ glm_mat3_mul(mat3 m1, mat3 m2, mat3 dest) {
dest[0][1] = a01 * b00 + a11 * b01 + a21 * b02;
dest[0][2] = a02 * b00 + a12 * b01 + a22 * b02;
dest[1][0] = a00 * b10 + a10 * b11 + a20 * b12;
dest[1][1] = a01 * b10 + a11 * b11 + a21 * b12;
dest[1][2] = a02 * b10 + a12 * b11 + a22 * b12;
dest[2][0] = a00 * b20 + a10 * b21 + a20 * b22;
dest[2][1] = a01 * b20 + a11 * b21 + a21 * b22;
dest[2][2] = a02 * b20 + a12 * b21 + a22 * b22;
#endif
}
@@ -87,21 +85,15 @@ glm_mat3_mul(mat3 m1, mat3 m2, mat3 dest) {
CGLM_INLINE
void
glm_mat3_transpose_to(mat3 m, mat3 dest) {
#if defined( __SSE__ ) || defined( __SSE2__ )
glm_mat3_transp_to_sse2(m, dest);
#else
dest[0][0] = m[0][0];
dest[0][1] = m[1][0];
dest[0][2] = m[2][0];
dest[1][0] = m[0][1];
dest[1][1] = m[1][1];
dest[1][2] = m[2][1];
dest[2][0] = m[0][2];
dest[2][1] = m[1][2];
dest[2][2] = m[2][2];
#endif
}
/*!
@@ -112,9 +104,6 @@ glm_mat3_transpose_to(mat3 m, mat3 dest) {
CGLM_INLINE
void
glm_mat3_transpose(mat3 m) {
#if defined( __SSE__ ) || defined( __SSE2__ )
glm_mat3_transp_sse2(m);
#else
mat3 tmp;
tmp[0][1] = m[1][0];
@@ -130,7 +119,6 @@ glm_mat3_transpose(mat3 m) {
m[1][2] = tmp[1][2];
m[2][0] = tmp[2][0];
m[2][1] = tmp[2][1];
#endif
}
/*!