mewin/glm
1
0
Fork 0
Code Releases Activity

Added C++17 detection, added VC language version detection, clean up alignment type, clean up swizzle operator detection

Browse Source
This commit is contained in:
Groove
2018-07-11 02:53:22 +02:00
parent bd2c7070d7
commit 16106b2e3e
21 changed files with 246 additions and 387 deletions
Download Patch File Download Diff File Expand all files Collapse all files

26
glm/simd/common.h
View File

@@ -7,52 +7,52 @@
#if GLM_ARCH & GLM_ARCH_SSE2_BIT
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_add(glm_vec4 a, glm_vec4 b)
GLM_FUNC_QUALIFIER glm_f32vec4 glm_vec4_add(glm_f32vec4 a, glm_f32vec4 b)
{
return _mm_add_ps(a, b);
}
GLM_FUNC_QUALIFIER glm_vec4 glm_vec1_add(glm_vec4 a, glm_vec4 b)
GLM_FUNC_QUALIFIER glm_f32vec4 glm_vec1_add(glm_f32vec4 a, glm_f32vec4 b)
{
return _mm_add_ss(a, b);
}
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_sub(glm_vec4 a, glm_vec4 b)
GLM_FUNC_QUALIFIER glm_f32vec4 glm_vec4_sub(glm_f32vec4 a, glm_f32vec4 b)
{
return _mm_sub_ps(a, b);
}
GLM_FUNC_QUALIFIER glm_vec4 glm_vec1_sub(glm_vec4 a, glm_vec4 b)
GLM_FUNC_QUALIFIER glm_f32vec4 glm_vec1_sub(glm_f32vec4 a, glm_f32vec4 b)
{
return _mm_sub_ss(a, b);
}
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_mul(glm_vec4 a, glm_vec4 b)
GLM_FUNC_QUALIFIER glm_f32vec4 glm_vec4_mul(glm_f32vec4 a, glm_f32vec4 b)
{
return _mm_mul_ps(a, b);
}
GLM_FUNC_QUALIFIER glm_vec4 glm_vec1_mul(glm_vec4 a, glm_vec4 b)
GLM_FUNC_QUALIFIER glm_f32vec4 glm_vec1_mul(glm_f32vec4 a, glm_f32vec4 b)
{
return _mm_mul_ss(a, b);
}
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_div(glm_vec4 a, glm_vec4 b)
GLM_FUNC_QUALIFIER glm_f32vec4 glm_vec4_div(glm_f32vec4 a, glm_f32vec4 b)
{
return _mm_div_ps(a, b);
}
GLM_FUNC_QUALIFIER glm_vec4 glm_vec1_div(glm_vec4 a, glm_vec4 b)
GLM_FUNC_QUALIFIER glm_f32vec4 glm_vec1_div(glm_f32vec4 a, glm_f32vec4 b)
{
return _mm_div_ss(a, b);
}
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_div_lowp(glm_vec4 a, glm_vec4 b)
GLM_FUNC_QUALIFIER glm_f32vec4 glm_vec4_div_lowp(glm_f32vec4 a, glm_f32vec4 b)
{
return glm_vec4_mul(a, _mm_rcp_ps(b));
}
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_swizzle_xyzw(glm_vec4 a)
GLM_FUNC_QUALIFIER glm_f32vec4 glm_vec4_swizzle_xyzw(glm_f32vec4 a)
{
# if GLM_ARCH & GLM_ARCH_AVX2_BIT
return _mm_permute_ps(a, _MM_SHUFFLE(3, 2, 1, 0));
@@ -61,7 +61,7 @@ GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_swizzle_xyzw(glm_vec4 a)
# endif
}
GLM_FUNC_QUALIFIER glm_vec4 glm_vec1_fma(glm_vec4 a, glm_vec4 b, glm_vec4 c)
GLM_FUNC_QUALIFIER glm_f32vec4 glm_vec1_fma(glm_f32vec4 a, glm_f32vec4 b, glm_f32vec4 c)
{
# if (GLM_ARCH & GLM_ARCH_AVX2_BIT) && !(GLM_COMPILER & GLM_COMPILER_CLANG)
return _mm_fmadd_ss(a, b, c);
@@ -70,7 +70,7 @@ GLM_FUNC_QUALIFIER glm_vec4 glm_vec1_fma(glm_vec4 a, glm_vec4 b, glm_vec4 c)
# endif
}
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_fma(glm_vec4 a, glm_vec4 b, glm_vec4 c)
GLM_FUNC_QUALIFIER glm_f32vec4 glm_vec4_fma(glm_f32vec4 a, glm_f32vec4 b, glm_f32vec4 c)
{
# if (GLM_ARCH & GLM_ARCH_AVX2_BIT) && !(GLM_COMPILER & GLM_COMPILER_CLANG)
return _mm_fmadd_ps(a, b, c);
@@ -79,7 +79,7 @@ GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_fma(glm_vec4 a, glm_vec4 b, glm_vec4 c)
# endif
}
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_abs(glm_vec4 x)
GLM_FUNC_QUALIFIER glm_f32vec4 glm_vec4_abs(glm_f32vec4 x)
{
return _mm_and_ps(x, _mm_castsi128_ps(_mm_set1_epi32(0x7FFFFFFF)));
}

4
glm/simd/exponential.h
View File

@@ -7,12 +7,12 @@
#if GLM_ARCH & GLM_ARCH_SSE2_BIT
GLM_FUNC_QUALIFIER glm_vec4 glm_vec1_sqrt_lowp(glm_vec4 x)
GLM_FUNC_QUALIFIER glm_f32vec4 glm_vec1_sqrt_lowp(glm_f32vec4 x)
{
return _mm_mul_ss(_mm_rsqrt_ss(x), x);
}
GLM_FUNC_QUALIFIER glm_vec4 glm_vec4_sqrt_lowp(glm_vec4 x)
GLM_FUNC_QUALIFIER glm_f32vec4 glm_vec4_sqrt_lowp(glm_f32vec4 x)
{
return _mm_mul_ps(_mm_rsqrt_ps(x), x);
}

21
glm/simd/platform.h
View File

@@ -385,16 +385,25 @@
#endif//GLM_ARCH
#if GLM_ARCH & GLM_ARCH_SSE2_BIT
typedef __m128 glm_vec4;
typedef __m128i glm_ivec4;
typedef __m128i glm_uvec4;
typedef __m128 glm_f32vec4;
typedef __m128i glm_i32vec4;
typedef __m128i glm_u32vec4;
typedef __m128d glm_f64vec2;
typedef __m128i glm_i64vec2;
typedef __m128i glm_u64vec2;
typedef glm_f32vec4 glm_vec4;
typedef glm_i32vec4 glm_ivec4;
typedef glm_u32vec4 glm_uvec4;
typedef glm_f64vec2 glm_dvec2;
#endif
#if GLM_ARCH & GLM_ARCH_AVX_BIT
typedef __m256d glm_dvec4;
typedef __m256d glm_f64vec4;
typedef glm_f64vec4 glm_dvec4;
#endif
#if GLM_ARCH & GLM_ARCH_AVX2_BIT
typedef __m256i glm_i64vec4;
typedef __m256i glm_u64vec4;
typedef __m256i glm_i64vec4;
typedef __m256i glm_u64vec4;
#endif