External/cglm
1
0
Fork 0
mirror of https://github.com/recp/cglm.git synced 2025-10-04 01:00:46 +00:00
Code Issues Packages Projects Releases Wiki Activity

optimize affine translate with SSE

Browse Source
This commit is contained in:
Recep Aslantas
2016-09-13 12:17:25 +03:00
parent 7ef105bab0
commit c480267463
1 changed files with 31 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

37
include/cglm-affine.h
View File

@@ -15,11 +15,25 @@
CGLM_INLINE CGLM_INLINE
void void
glm_translate_to(mat4 m, vec3 v, mat4 dest) { glm_translate_to(mat4 m, vec3 v, mat4 dest) {
vec4 v1;
vec4 v2;
vec4 v3;
mat4 t = GLM_MAT_IDENTITY_4F; mat4 t = GLM_MAT_IDENTITY_4F;
#if defined( __SSE__ ) || defined( __SSE2__ )
_mm_store_ps(dest[3],
_mm_add_ps(_mm_add_ps(_mm_mul_ps(_mm_load_ps(t[0]),
_mm_set1_ps(v[0])),
_mm_mul_ps(_mm_load_ps(t[1]),
_mm_set1_ps(v[1]))),
_mm_add_ps(_mm_mul_ps(_mm_load_ps(t[2]),
_mm_set1_ps(v[2])),
_mm_load_ps(t[3]))))
;
_mm_store_ps(dest[0], _mm_load_ps(m[0]));
_mm_store_ps(dest[1], _mm_load_ps(m[1]));
_mm_store_ps(dest[2], _mm_load_ps(m[2]));
#else
vec4 v1, v2, v3;
glm_vec4_scale(t[0], v[0], v1); glm_vec4_scale(t[0], v[0], v1);
glm_vec4_scale(t[1], v[1], v2); glm_vec4_scale(t[1], v[1], v2);
glm_vec4_scale(t[2], v[2], v3); glm_vec4_scale(t[2], v[2], v3);
@@ -29,14 +43,24 @@ glm_translate_to(mat4 m, vec3 v, mat4 dest) {
glm_vec4_add(v3, t[3], t[3]); glm_vec4_add(v3, t[3], t[3]);
glm__memcpy(float, dest, t, sizeof(mat4)); glm__memcpy(float, dest, t, sizeof(mat4));
#endif
} }
CGLM_INLINE CGLM_INLINE
void void
glm_translate(mat4 m, vec3 v) { glm_translate(mat4 m, vec3 v) {
vec4 v1; #if defined( __SSE__ ) || defined( __SSE2__ )
vec4 v2; _mm_store_ps(m[3],
vec4 v3; _mm_add_ps(_mm_add_ps(_mm_mul_ps(_mm_load_ps(m[0]),
_mm_set1_ps(v[0])),
_mm_mul_ps(_mm_load_ps(m[1]),
_mm_set1_ps(v[1]))),
_mm_add_ps(_mm_mul_ps(_mm_load_ps(m[2]),
_mm_set1_ps(v[2])),
_mm_load_ps(m[3]))))
;
#else
vec4 v1, v2, v3;
glm_vec4_scale(m[0], v[0], v1); glm_vec4_scale(m[0], v[0], v1);
glm_vec4_scale(m[1], v[1], v2); glm_vec4_scale(m[1], v[1], v2);
@@ -45,6 +69,7 @@ glm_translate(mat4 m, vec3 v) {
glm_vec4_add(v1, m[3], m[3]); glm_vec4_add(v1, m[3], m[3]);
glm_vec4_add(v2, m[3], m[3]); glm_vec4_add(v2, m[3], m[3]);
glm_vec4_add(v3, m[3], m[3]); glm_vec4_add(v3, m[3], m[3]);
#endif
} }
CGLM_INLINE CGLM_INLINE