mirror/glm
1
0
Fork 0
mirror of https://github.com/g-truc/glm.git synced 2024-11-22 08:54:35 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #1049 from borodust/aligned_bugfixes

Browse Source 
Various bugfixes for aligned/simd codepaths #1049
This commit is contained in:
Christophe 2022-04-20 12:39:39 +02:00 committed by GitHub
commit cc98465e35
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
3 changed files with 9 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
glm/detail/func_matrix_simd.inl
View File

@ -18,9 +18,9 @@ namespace detail
{ {
mat<4, 4, float, Q> Result; mat<4, 4, float, Q> Result;
glm_mat4_matrixCompMult( glm_mat4_matrixCompMult(
*static_cast<glm_vec4 const (*)[4]>(&x[0].data), &x[0].data,
*static_cast<glm_vec4 const (*)[4]>(&y[0].data), &y[0].data,
*static_cast<glm_vec4(*)[4]>(&Result[0].data)); &Result[0].data);
return Result; return Result;
} }
}; };

1
glm/detail/qualifier.hpp
View File

@ -144,7 +144,6 @@ namespace detail
typedef glm_u64vec2 type; typedef glm_u64vec2 type;
}; };
# endif # endif
# if (GLM_ARCH & GLM_ARCH_AVX_BIT) # if (GLM_ARCH & GLM_ARCH_AVX_BIT)
template<> template<>
struct storage<4, double, true> struct storage<4, double, true>

13
glm/detail/type_quat_simd.inl
View File

@ -14,10 +14,10 @@ namespace detail
// SSE2 STATS: 11 shuffle, 8 mul, 8 add // SSE2 STATS: 11 shuffle, 8 mul, 8 add
// SSE4 STATS: 3 shuffle, 4 mul, 4 dpps // SSE4 STATS: 3 shuffle, 4 mul, 4 dpps
__m128 const mul0 = _mm_mul_ps(q1.Data, _mm_shuffle_ps(q2.Data, q2.Data, _MM_SHUFFLE(0, 1, 2, 3))); __m128 const mul0 = _mm_mul_ps(q1.data, _mm_shuffle_ps(q2.data, q2.data, _MM_SHUFFLE(0, 1, 2, 3)));
__m128 const mul1 = _mm_mul_ps(q1.Data, _mm_shuffle_ps(q2.Data, q2.Data, _MM_SHUFFLE(1, 0, 3, 2))); __m128 const mul1 = _mm_mul_ps(q1.data, _mm_shuffle_ps(q2.data, q2.data, _MM_SHUFFLE(1, 0, 3, 2)));
__m128 const mul2 = _mm_mul_ps(q1.Data, _mm_shuffle_ps(q2.Data, q2.Data, _MM_SHUFFLE(2, 3, 0, 1))); __m128 const mul2 = _mm_mul_ps(q1.data, _mm_shuffle_ps(q2.data, q2.data, _MM_SHUFFLE(2, 3, 0, 1)));
__m128 const mul3 = _mm_mul_ps(q1.Data, q2.Data); __m128 const mul3 = _mm_mul_ps(q1.data, q2.data);
# if GLM_ARCH & GLM_ARCH_SSE41_BIT # if GLM_ARCH & GLM_ARCH_SSE41_BIT
__m128 const add0 = _mm_dp_ps(mul0, _mm_set_ps(1.0f, -1.0f, 1.0f, 1.0f), 0xff); __m128 const add0 = _mm_dp_ps(mul0, _mm_set_ps(1.0f, -1.0f, 1.0f, 1.0f), 0xff);
@ -89,7 +89,7 @@ namespace detail
{ {
static qua<float, Q> call(qua<float, Q> const& q, qua<float, Q> const& p) static qua<float, Q> call(qua<float, Q> const& q, qua<float, Q> const& p)
{ {
vec<4, float, Q> Result; qua<float, Q> Result;
Result.data = _mm_sub_ps(q.data, p.data); Result.data = _mm_sub_ps(q.data, p.data);
return Result; return Result;
} }
@ -177,7 +177,7 @@ namespace detail
uuv = _mm_mul_ps(uuv, two); uuv = _mm_mul_ps(uuv, two);
vec<4, float, Q> Result; vec<4, float, Q> Result;
Result.data = _mm_add_ps(v.Data, _mm_add_ps(uv, uuv)); Result.data = _mm_add_ps(v.data, _mm_add_ps(uv, uuv));
return Result; return Result;
} }
}; };
@ -185,4 +185,3 @@ namespace detail
}//namespace glm }//namespace glm
#endif//GLM_ARCH & GLM_ARCH_SSE2_BIT #endif//GLM_ARCH & GLM_ARCH_SSE2_BIT