External/cglm
1
0
Fork 0
mirror of https://github.com/recp/cglm.git synced 2025-10-03 16:51:35 +00:00
Code Issues Packages Projects Releases Wiki Activity

arm, neon: neon/fma support for glm_mul_rot()

Browse Source
This commit is contained in:
Recep Aslantas
2021-04-25 03:49:35 +03:00
parent 2903813765
commit 155eb109a8
2 changed files with 31 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
include/cglm/affine-mat.h
View File

@@ -109,6 +109,8 @@ void
glm_mul_rot(mat4 m1, mat4 m2, mat4 dest) {
#if defined( __SSE__ ) || defined( __SSE2__ )
glm_mul_rot_sse2(m1, m2, dest);
#elif defined(CGLM_NEON_FP)
glm_mul_rot_neon(m1, m2, dest);
#else
float a00 = m1[0][0], a01 = m1[0][1], a02 = m1[0][2], a03 = m1[0][3],
a10 = m1[1][0], a11 = m1[1][1], a12 = m1[1][2], a13 = m1[1][3],

29
include/cglm/simd/neon/affine.h
View File

@@ -47,5 +47,34 @@ glm_mul_neon(mat4 m1, mat4 m2, mat4 dest) {
glmm_store(dest[3], v3);
}
CGLM_INLINE
void
glm_mul_rot_neon(mat4 m1, mat4 m2, mat4 dest) {
/* D = R * L (Column-Major) */
glmm_128 l0, l1, l2, r0, r1, r2, v0, v1, v2;
l0 = glmm_load(m1[0]); r0 = glmm_load(m2[0]);
l1 = glmm_load(m1[1]); r1 = glmm_load(m2[1]);
l2 = glmm_load(m1[2]); r2 = glmm_load(m2[2]);
v0 = vmulq_f32(glmm_splat_x(r0), l0);
v1 = vmulq_f32(glmm_splat_x(r1), l0);
v2 = vmulq_f32(glmm_splat_x(r2), l0);
v0 = glmm_fmadd(glmm_splat_y(r0), l1, v0);
v1 = glmm_fmadd(glmm_splat_y(r1), l1, v1);
v2 = glmm_fmadd(glmm_splat_y(r2), l1, v2);
v0 = glmm_fmadd(glmm_splat_z(r0), l2, v0);
v1 = glmm_fmadd(glmm_splat_z(r1), l2, v1);
v2 = glmm_fmadd(glmm_splat_z(r2), l2, v2);
glmm_store(dest[0], v0);
glmm_store(dest[1], v1);
glmm_store(dest[2], v2);
glmm_store(dest[3], glmm_load(m1[3]));
}
#endif
#endif /* cglm_affine_neon_h */