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

optimize vec4 with SIMD (SSE)

Browse Source
This commit is contained in:
Recep Aslantas
2016-09-13 11:54:51 +03:00
parent b957509c9e
commit 705a4bb9a0
2 changed files with 53 additions and 30 deletions
Download Patch File Download Diff File Expand all files Collapse all files

44
include/cglm-affine.h
View File

@@ -18,13 +18,13 @@ glm_translate_to(mat4 m, vec3 v, mat4 dest) {
vec4 v3; vec4 v3;
mat4 t = GLM_MAT_IDENTITY_4F; mat4 t = GLM_MAT_IDENTITY_4F;
glm_vec_scale4(t[0], v[0], v1); glm_vec4_scale(t[0], v[0], v1);
glm_vec_scale4(t[1], v[1], v2); glm_vec4_scale(t[1], v[1], v2);
glm_vec_scale4(t[2], v[2], v3); glm_vec4_scale(t[2], v[2], v3);
glm_vec_add4(v1, t[3], t[3]); glm_vec4_add(v1, t[3], t[3]);
glm_vec_add4(v2, t[3], t[3]); glm_vec4_add(v2, t[3], t[3]);
glm_vec_add4(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));
} }
@@ -36,37 +36,37 @@ glm_translate(mat4 m, vec3 v) {
vec4 v2; vec4 v2;
vec4 v3; vec4 v3;
glm_vec_scale4(m[0], v[0], v1); glm_vec4_scale(m[0], v[0], v1);
glm_vec_scale4(m[1], v[1], v2); glm_vec4_scale(m[1], v[1], v2);
glm_vec_scale4(m[2], v[2], v3); glm_vec4_scale(m[2], v[2], v3);
glm_vec_add4(v1, m[3], m[3]); glm_vec4_add(v1, m[3], m[3]);
glm_vec_add4(v2, m[3], m[3]); glm_vec4_add(v2, m[3], m[3]);
glm_vec_add4(v3, m[3], m[3]); glm_vec4_add(v3, m[3], m[3]);
} }
CGLM_INLINE CGLM_INLINE
void void
glm_translate_x(mat4 m, float to) { glm_translate_x(mat4 m, float to) {
vec4 v1; vec4 v1;
glm_vec_scale4(m[0], to, v1); glm_vec4_scale(m[0], to, v1);
glm_vec_add4(v1, m[3], m[3]); glm_vec4_add(v1, m[3], m[3]);
} }
CGLM_INLINE CGLM_INLINE
void void
glm_translate_y(mat4 m, float to) { glm_translate_y(mat4 m, float to) {
vec4 v1; vec4 v1;
glm_vec_scale4(m[1], to, v1); glm_vec4_scale(m[1], to, v1);
glm_vec_add4(v1, m[3], m[3]); glm_vec4_add(v1, m[3], m[3]);
} }
CGLM_INLINE CGLM_INLINE
void void
glm_translate_z(mat4 m, float to) { glm_translate_z(mat4 m, float to) {
vec4 v1; vec4 v1;
glm_vec_scale4(m[2], to, v1); glm_vec4_scale(m[2], to, v1);
glm_vec_add4(v1, m[3], m[3]); glm_vec4_add(v1, m[3], m[3]);
} }
/* scale */ /* scale */
@@ -74,11 +74,11 @@ glm_translate_z(mat4 m, float to) {
CGLM_INLINE CGLM_INLINE
void void
glm_scale_to(mat4 m, vec3 v, mat4 dest) { glm_scale_to(mat4 m, vec3 v, mat4 dest) {
glm_vec_scale4(m[0], v[0], dest[0]); glm_vec4_scale(m[0], v[0], dest[0]);
glm_vec_scale4(m[1], v[1], dest[1]); glm_vec4_scale(m[1], v[1], dest[1]);
glm_vec_scale4(m[2], v[2], dest[2]); glm_vec4_scale(m[2], v[2], dest[2]);
glm_vec_dup4(m[3], dest[3]); glm_vec4_dup(m[3], dest[3]);
} }
CGLM_INLINE CGLM_INLINE

39
include/cglm-vec.h
View File

@@ -9,6 +9,7 @@
#define cglm_vec_h #define cglm_vec_h
#include "cglm.h" #include "cglm.h"
#include "cglm-intrin.h"
CGLM_INLINE CGLM_INLINE
void void
@@ -20,11 +21,15 @@ glm_vec_dup(vec3 a, vec3 dest) {
CGLM_INLINE CGLM_INLINE
void void
glm_vec_dup4(vec4 a, vec4 dest) { glm_vec4_dup(vec4 v, vec4 dest) {
dest[0] = a[0]; #if defined( __SSE__ ) || defined( __SSE2__ )
dest[1] = a[1]; _mm_store_ps(dest, _mm_load_ps(v));
dest[2] = a[2]; #else
dest[3] = a[3]; dest[0] = v[0];
dest[1] = v[1];
dest[2] = v[2];
dest[3] = v[3];
#endif
} }
CGLM_INLINE CGLM_INLINE
@@ -98,11 +103,17 @@ glm_vec_add(vec3 v1, vec3 v2, vec3 dest) {
CGLM_INLINE CGLM_INLINE
void void
glm_vec_add4(vec4 v1, vec4 v2, vec4 dest) { glm_vec4_add(vec4 v1, vec4 v2, vec4 dest) {
#if defined( __SSE__ ) || defined( __SSE2__ )
_mm_store_ps(dest,
_mm_add_ps(_mm_load_ps(v1),
_mm_load_ps(v2)));
#else
dest[0] = v1[0] + v2[0]; dest[0] = v1[0] + v2[0];
dest[1] = v1[1] + v2[1]; dest[1] = v1[1] + v2[1];
dest[2] = v1[2] + v2[2]; dest[2] = v1[2] + v2[2];
dest[3] = v1[3] + v2[3]; dest[3] = v1[3] + v2[3];
#endif
} }
CGLM_INLINE CGLM_INLINE
@@ -115,11 +126,17 @@ glm_vec_sub(vec3 v1, vec3 v2, vec3 dest) {
CGLM_INLINE CGLM_INLINE
void void
glm_vec_sub4(vec4 v1, vec4 v2, vec4 dest) { glm_vec4_sub(vec4 v1, vec4 v2, vec4 dest) {
#if defined( __SSE__ ) || defined( __SSE2__ )
_mm_store_ps(dest,
_mm_sub_ps(_mm_load_ps(v1),
_mm_load_ps(v2)));
#else
dest[0] = v1[0] - v2[0]; dest[0] = v1[0] - v2[0];
dest[1] = v1[1] - v2[1]; dest[1] = v1[1] - v2[1];
dest[2] = v1[2] - v2[2]; dest[2] = v1[2] - v2[2];
dest[3] = v1[3] - v2[3]; dest[3] = v1[3] - v2[3];
#endif
} }
CGLM_INLINE CGLM_INLINE
@@ -132,11 +149,17 @@ glm_vec_scale(vec3 v, float s, vec3 dest) {
CGLM_INLINE CGLM_INLINE
void void
glm_vec_scale4(vec3 v, float s, vec3 dest) { glm_vec4_scale(vec4 v, float s, vec4 dest) {
#if defined( __SSE__ ) || defined( __SSE2__ )
_mm_store_ps(dest,
_mm_mul_ps(_mm_load_ps(v),
_mm_set1_ps(s)));
#else
dest[0] = v[0] * s; dest[0] = v[0] * s;
dest[1] = v[1] * s; dest[1] = v[1] * s;
dest[2] = v[2] * s; dest[2] = v[2] * s;
dest[3] = v[3] * s; dest[3] = v[3] * s;
#endif
} }
CGLM_INLINE CGLM_INLINE