mirror/glm
1
0
Fork 0
mirror of https://github.com/g-truc/glm.git synced 2024-11-10 12:41:54 +00:00
Code Issues Packages Projects Releases Wiki Activity

Renamed instruction set flags

Browse Source
This commit is contained in:
Christophe Riccio 2016-05-28 19:26:59 +02:00
parent d33e3df02a
commit 7fe2f5fe65
19 changed files with 122 additions and 123 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
glm/detail/func_common_simd.inl
View File

@ -1,7 +1,7 @@
/// @ref core /// @ref core
/// @file glm/detail/func_common_simd.inl /// @file glm/detail/func_common_simd.inl
#if GLM_ARCH & GLM_ARCH_SSE2_FLAG #if GLM_ARCH & GLM_ARCH_SSE2_BIT
#include "../simd/common.h" #include "../simd/common.h"
@ -135,4 +135,4 @@ namespace detail
}//namespace detail }//namespace detail
}//namespace glm }//namespace glm
#endif//GLM_ARCH & GLM_ARCH_SSE2_FLAG #endif//GLM_ARCH & GLM_ARCH_SSE2_BIT

4
glm/detail/func_geometric_simd.inl
View File

@ -1,6 +1,6 @@
#include "../simd/geometric.h" #include "../simd/geometric.h"
#if GLM_ARCH & GLM_ARCH_SSE2_FLAG #if GLM_ARCH & GLM_ARCH_SSE2_BIT
namespace glm{ namespace glm{
namespace detail namespace detail
@ -17,4 +17,4 @@ namespace detail
}//namespace detail }//namespace detail
}//namespace glm }//namespace glm
#endif//GLM_ARCH & GLM_ARCH_SSE2_FLAG #endif//GLM_ARCH & GLM_ARCH_SSE2_BIT

2
glm/detail/func_matrix_simd.inl
View File

@ -8,7 +8,7 @@
namespace glm{ namespace glm{
namespace detail namespace detail
{ {
# if GLM_ARCH & GLM_ARCH_SSE2_FLAG # if GLM_ARCH & GLM_ARCH_SSE2_BIT
template <precision P> template <precision P>
struct compute_inverse<tmat4x4, float, P> struct compute_inverse<tmat4x4, float, P>
{ {

6
glm/detail/type_vec4.hpp
View File

@ -29,7 +29,7 @@ namespace detail
typedef T type[4]; typedef T type[4];
}; };
# if (GLM_ARCH & GLM_ARCH_SSE2_FLAG) # if (GLM_ARCH & GLM_ARCH_SSE2_BIT)
template <> template <>
struct simd_data<float> struct simd_data<float>
{ {
@ -49,7 +49,7 @@ namespace detail
}; };
# endif # endif
# if (GLM_ARCH & GLM_ARCH_AVX_FLAG) # if (GLM_ARCH & GLM_ARCH_AVX_BIT)
template <> template <>
struct simd_data<double> struct simd_data<double>
{ {
@ -57,7 +57,7 @@ namespace detail
}; };
# endif # endif
# if (GLM_ARCH & GLM_ARCH_AVX2_FLAG) # if (GLM_ARCH & GLM_ARCH_AVX2_BIT)
template <> template <>
struct simd_data<int64> struct simd_data<int64>
{ {

20
glm/detail/type_vec4_simd.inl
View File

@ -1,7 +1,7 @@
/// @ref core /// @ref core
/// @file glm/detail/type_tvec4_simd.inl /// @file glm/detail/type_tvec4_simd.inl
#if GLM_ARCH & GLM_ARCH_SSE2_FLAG #if GLM_ARCH & GLM_ARCH_SSE2_BIT
namespace glm{ namespace glm{
namespace detail namespace detail
@ -72,7 +72,7 @@ namespace detail
} }
}; };
# if GLM_ARCH & GLM_ARCH_AVX2_FLAG # if GLM_ARCH & GLM_ARCH_AVX2_BIT
template <typename T, precision P> template <typename T, precision P>
struct compute_vec4_and<T, P, true, 64> struct compute_vec4_and<T, P, true, 64>
{ {
@ -96,7 +96,7 @@ namespace detail
} }
}; };
# if GLM_ARCH & GLM_ARCH_AVX2_FLAG # if GLM_ARCH & GLM_ARCH_AVX2_BIT
template <typename T, precision P> template <typename T, precision P>
struct compute_vec4_or<T, P, true, 64> struct compute_vec4_or<T, P, true, 64>
{ {
@ -120,7 +120,7 @@ namespace detail
} }
}; };
# if GLM_ARCH & GLM_ARCH_AVX2_FLAG # if GLM_ARCH & GLM_ARCH_AVX2_BIT
template <typename T, precision P> template <typename T, precision P>
struct compute_vec4_xor<T, P, true, 64> struct compute_vec4_xor<T, P, true, 64>
{ {
@ -144,7 +144,7 @@ namespace detail
} }
}; };
# if GLM_ARCH & GLM_ARCH_AVX2_FLAG # if GLM_ARCH & GLM_ARCH_AVX2_BIT
template <typename T, precision P> template <typename T, precision P>
struct compute_vec4_shift_left<T, P, true, 64> struct compute_vec4_shift_left<T, P, true, 64>
{ {
@ -168,7 +168,7 @@ namespace detail
} }
}; };
# if GLM_ARCH & GLM_ARCH_AVX2_FLAG # if GLM_ARCH & GLM_ARCH_AVX2_BIT
template <typename T, precision P> template <typename T, precision P>
struct compute_vec4_shift_right<T, P, true, 64> struct compute_vec4_shift_right<T, P, true, 64>
{ {
@ -192,7 +192,7 @@ namespace detail
} }
}; };
# if GLM_ARCH & GLM_ARCH_AVX2_FLAG # if GLM_ARCH & GLM_ARCH_AVX2_BIT
template <typename T, precision P> template <typename T, precision P>
struct compute_vec4_bitwise_not<T, P, true, 64> struct compute_vec4_bitwise_not<T, P, true, 64>
{ {
@ -248,7 +248,7 @@ namespace detail
data(_mm_set1_ps(s)) data(_mm_set1_ps(s))
{} {}
# if GLM_ARCH & GLM_ARCH_AVX_FLAG # if GLM_ARCH & GLM_ARCH_AVX_BIT
template <> template <>
GLM_FUNC_QUALIFIER GLM_CONSTEXPR_SIMD tvec4<double, lowp>::tvec4(double s) : GLM_FUNC_QUALIFIER GLM_CONSTEXPR_SIMD tvec4<double, lowp>::tvec4(double s) :
data(_mm256_set1_pd(s)) data(_mm256_set1_pd(s))
@ -280,7 +280,7 @@ namespace detail
data(_mm_set1_epi32(s)) data(_mm_set1_epi32(s))
{} {}
# if GLM_ARCH & GLM_ARCH_AVX2_FLAG # if GLM_ARCH & GLM_ARCH_AVX2_BIT
template <> template <>
GLM_FUNC_QUALIFIER GLM_CONSTEXPR_SIMD tvec4<int64, lowp>::tvec4(int64 s) : GLM_FUNC_QUALIFIER GLM_CONSTEXPR_SIMD tvec4<int64, lowp>::tvec4(int64 s) :
data(_mm256_set1_epi64x(s)) data(_mm256_set1_epi64x(s))
@ -350,4 +350,4 @@ namespace detail
*/ */
}//namespace glm }//namespace glm
#endif//GLM_ARCH & GLM_ARCH_SSE2_FLAG #endif//GLM_ARCH & GLM_ARCH_SSE2_BIT

2
glm/gtx/simd_mat4.hpp
View File

@ -17,7 +17,7 @@
#if(GLM_ARCH != GLM_ARCH_PURE) #if(GLM_ARCH != GLM_ARCH_PURE)
#if(GLM_ARCH & GLM_ARCH_SSE2_FLAG) #if(GLM_ARCH & GLM_ARCH_SSE2_BIT)
# include "../detail/intrinsic_matrix.hpp" # include "../detail/intrinsic_matrix.hpp"
# include "../gtx/simd_vec4.hpp" # include "../gtx/simd_vec4.hpp"
#else #else

2
glm/gtx/simd_quat.hpp
View File

@ -19,7 +19,7 @@
#if(GLM_ARCH != GLM_ARCH_PURE) #if(GLM_ARCH != GLM_ARCH_PURE)
#if(GLM_ARCH & GLM_ARCH_SSE2_FLAG) #if(GLM_ARCH & GLM_ARCH_SSE2_BIT)
# include "../gtx/simd_mat4.hpp" # include "../gtx/simd_mat4.hpp"
#else #else
# error "GLM: GLM_GTX_simd_quat requires compiler support of SSE2 through intrinsics" # error "GLM: GLM_GTX_simd_quat requires compiler support of SSE2 through intrinsics"

2
glm/gtx/simd_quat.inl
View File

@ -122,7 +122,7 @@ GLM_FUNC_QUALIFIER fquatSIMD operator* (fquatSIMD const & q1, fquatSIMD const &
__m128 mul2 = _mm_mul_ps(q1.Data, _mm_shuffle_ps(q2.Data, q2.Data, _MM_SHUFFLE(2, 3, 0, 1))); __m128 mul2 = _mm_mul_ps(q1.Data, _mm_shuffle_ps(q2.Data, q2.Data, _MM_SHUFFLE(2, 3, 0, 1)));
__m128 mul3 = _mm_mul_ps(q1.Data, q2.Data); __m128 mul3 = _mm_mul_ps(q1.Data, q2.Data);
# if(GLM_ARCH & GLM_ARCH_SSE41_FLAG) # if(GLM_ARCH & GLM_ARCH_SSE41_BIT)
__m128 add0 = _mm_dp_ps(mul0, _mm_set_ps(1.0f, -1.0f, 1.0f, 1.0f), 0xff); __m128 add0 = _mm_dp_ps(mul0, _mm_set_ps(1.0f, -1.0f, 1.0f, 1.0f), 0xff);
__m128 add1 = _mm_dp_ps(mul1, _mm_set_ps(1.0f, 1.0f, 1.0f, -1.0f), 0xff); __m128 add1 = _mm_dp_ps(mul1, _mm_set_ps(1.0f, 1.0f, 1.0f, -1.0f), 0xff);
__m128 add2 = _mm_dp_ps(mul2, _mm_set_ps(1.0f, 1.0f, -1.0f, 1.0f), 0xff); __m128 add2 = _mm_dp_ps(mul2, _mm_set_ps(1.0f, 1.0f, -1.0f, 1.0f), 0xff);

2
glm/gtx/simd_vec4.hpp
View File

@ -17,7 +17,7 @@
#if(GLM_ARCH != GLM_ARCH_PURE) #if(GLM_ARCH != GLM_ARCH_PURE)
#if(GLM_ARCH & GLM_ARCH_SSE2_FLAG) #if(GLM_ARCH & GLM_ARCH_SSE2_BIT)
# include "../detail/intrinsic_common.hpp" # include "../detail/intrinsic_common.hpp"
# include "../detail/intrinsic_geometric.hpp" # include "../detail/intrinsic_geometric.hpp"
# include "../detail/intrinsic_integer.hpp" # include "../detail/intrinsic_integer.hpp"

10
glm/simd/common.h
View File

@ -5,12 +5,12 @@
#include "platform.h" #include "platform.h"
#if GLM_ARCH & GLM_ARCH_SSE2_FLAG #if GLM_ARCH & GLM_ARCH_SSE2_BIT
//mad //mad
GLM_FUNC_QUALIFIER __m128 glm_f32v1_mad(__m128 a, __m128 b, __m128 c) GLM_FUNC_QUALIFIER __m128 glm_f32v1_mad(__m128 a, __m128 b, __m128 c)
{ {
# if GLM_ARCH & GLM_ARCH_AVX2_FLAG # if GLM_ARCH & GLM_ARCH_AVX2_BIT
return _mm_fmadd_ss(a, b, c); return _mm_fmadd_ss(a, b, c);
# else # else
return _mm_add_ss(_mm_mul_ss(a, b), c); return _mm_add_ss(_mm_mul_ss(a, b), c);
@ -20,7 +20,7 @@ GLM_FUNC_QUALIFIER __m128 glm_f32v1_mad(__m128 a, __m128 b, __m128 c)
//mad //mad
GLM_FUNC_QUALIFIER __m128 glm_f32v4_mad(__m128 a, __m128 b, __m128 c) GLM_FUNC_QUALIFIER __m128 glm_f32v4_mad(__m128 a, __m128 b, __m128 c)
{ {
# if GLM_ARCH & GLM_ARCH_AVX2_FLAG # if GLM_ARCH & GLM_ARCH_AVX2_BIT
return _mm_fmadd_ps(a, b, c); return _mm_fmadd_ps(a, b, c);
# else # else
return _mm_add_ps(_mm_mul_ps(a, b), c); return _mm_add_ps(_mm_mul_ps(a, b), c);
@ -35,7 +35,7 @@ GLM_FUNC_QUALIFIER __m128 glm_f32v4_abs(__m128 x)
GLM_FUNC_QUALIFIER __m128i glm_i32v4_abs(__m128i x) GLM_FUNC_QUALIFIER __m128i glm_i32v4_abs(__m128i x)
{ {
# if GLM_ARCH & GLM_ARCH_SSSE3_FLAG # if GLM_ARCH & GLM_ARCH_SSSE3_BIT
return _mm_sign_epi32(x, x); return _mm_sign_epi32(x, x);
# else # else
__m128i const sgn0 = _mm_srai_epi32(x, 31); __m128i const sgn0 = _mm_srai_epi32(x, 31);
@ -204,4 +204,4 @@ GLM_FUNC_QUALIFIER __m128 glm_f32v4_sqrt_wip(__m128 x)
return Mul3; return Mul3;
} }
#endif//GLM_ARCH & GLM_ARCH_SSE2_FLAG #endif//GLM_ARCH & GLM_ARCH_SSE2_BIT

83
glm/simd/geometric.h
View File

@ -5,17 +5,17 @@
#include "common.h" #include "common.h"
#if GLM_ARCH & GLM_ARCH_SSE2_FLAG #if GLM_ARCH & GLM_ARCH_SSE2_BIT
GLM_FUNC_QUALIFIER __m128 glm_f32v4_dot(__m128 v1, __m128 v2) GLM_FUNC_QUALIFIER __m128 glm_f32v4_dot(__m128 v1, __m128 v2)
{ {
# if GLM_ARCH & GLM_ARCH_AVX_FLAG # if GLM_ARCH & GLM_ARCH_AVX_BIT
return _mm_dp_ps(v1, v2, 0xff); return _mm_dp_ps(v1, v2, 0xff);
# elif GLM_ARCH & GLM_ARCH_SSE3_FLAG # elif GLM_ARCH & GLM_ARCH_SSE3_BIT
__m128 const Mul0 = _mm_mul_ps(v1, v2); __m128 const mul0 = _mm_mul_ps(v1, v2);
__m128 const Hadd0 = _mm_hadd_ps(Mul0, Mul0); __m128 const hadd0 = _mm_hadd_ps(mul0, mul0);
__m128 const Hadd1 = _mm_hadd_ps(Hadd0, Hadd0); __m128 const hadd1 = _mm_hadd_ps(hadd0, hadd0);
return Hadd1; return hadd1;
# else # else
__m128 const mul0 = _mm_mul_ps(v1, v2); __m128 const mul0 = _mm_mul_ps(v1, v2);
__m128 const swp0 = _mm_shuffle_ps(mul0, mul0, _MM_SHUFFLE(2, 3, 0, 1)); __m128 const swp0 = _mm_shuffle_ps(mul0, mul0, _MM_SHUFFLE(2, 3, 0, 1));
@ -28,9 +28,9 @@ GLM_FUNC_QUALIFIER __m128 glm_f32v4_dot(__m128 v1, __m128 v2)
GLM_FUNC_QUALIFIER __m128 glm_f32v1_dot(__m128 v1, __m128 v2) GLM_FUNC_QUALIFIER __m128 glm_f32v1_dot(__m128 v1, __m128 v2)
{ {
# if GLM_ARCH & GLM_ARCH_AVX_FLAG # if GLM_ARCH & GLM_ARCH_AVX_BIT
return _mm_dp_ps(v1, v2, 0xff); return _mm_dp_ps(v1, v2, 0xff);
# elif GLM_ARCH & GLM_ARCH_SSE3_FLAG # elif GLM_ARCH & GLM_ARCH_SSE3_BIT
__m128 const mul0 = _mm_mul_ps(v1, v2); __m128 const mul0 = _mm_mul_ps(v1, v2);
__m128 const had0 = _mm_hadd_ps(mul0, mul0); __m128 const had0 = _mm_hadd_ps(mul0, mul0);
__m128 const had1 = _mm_hadd_ps(had0, had0); __m128 const had1 = _mm_hadd_ps(had0, had0);
@ -54,69 +54,68 @@ GLM_FUNC_QUALIFIER __m128 glm_f32v4_len(__m128 x)
GLM_FUNC_QUALIFIER __m128 glm_f32v4_dst(__m128 p0, __m128 p1) GLM_FUNC_QUALIFIER __m128 glm_f32v4_dst(__m128 p0, __m128 p1)
{ {
__m128 sub0 = _mm_sub_ps(p0, p1); __m128 const sub0 = _mm_sub_ps(p0, p1);
__m128 len0 = glm_f32v4_len(sub0); __m128 const len0 = glm_f32v4_len(sub0);
return len0; return len0;
} }
GLM_FUNC_QUALIFIER __m128 glm_f32v4_xpd(__m128 v1, __m128 v2) GLM_FUNC_QUALIFIER __m128 glm_f32v4_xpd(__m128 v1, __m128 v2)
{ {
__m128 swp0 = _mm_shuffle_ps(v1, v1, _MM_SHUFFLE(3, 0, 2, 1)); __m128 const swp0 = _mm_shuffle_ps(v1, v1, _MM_SHUFFLE(3, 0, 2, 1));
__m128 swp1 = _mm_shuffle_ps(v1, v1, _MM_SHUFFLE(3, 1, 0, 2)); __m128 const swp1 = _mm_shuffle_ps(v1, v1, _MM_SHUFFLE(3, 1, 0, 2));
__m128 swp2 = _mm_shuffle_ps(v2, v2, _MM_SHUFFLE(3, 0, 2, 1)); __m128 const swp2 = _mm_shuffle_ps(v2, v2, _MM_SHUFFLE(3, 0, 2, 1));
__m128 swp3 = _mm_shuffle_ps(v2, v2, _MM_SHUFFLE(3, 1, 0, 2)); __m128 const swp3 = _mm_shuffle_ps(v2, v2, _MM_SHUFFLE(3, 1, 0, 2));
__m128 mul0 = _mm_mul_ps(swp0, swp3); __m128 const mul0 = _mm_mul_ps(swp0, swp3);
__m128 mul1 = _mm_mul_ps(swp1, swp2); __m128 const mul1 = _mm_mul_ps(swp1, swp2);
__m128 sub0 = _mm_sub_ps(mul0, mul1); __m128 const sub0 = _mm_sub_ps(mul0, mul1);
return sub0; return sub0;
} }
GLM_FUNC_QUALIFIER __m128 glm_f32v4_nrm(__m128 v) GLM_FUNC_QUALIFIER __m128 glm_f32v4_nrm(__m128 v)
{ {
__m128 dot0 = glm_f32v4_dot(v, v); __m128 const dot0 = glm_f32v4_dot(v, v);
__m128 isr0 = _mm_rsqrt_ps(dot0); __m128 const isr0 = _mm_rsqrt_ps(dot0);
__m128 mul0 = _mm_mul_ps(v, isr0); __m128 const mul0 = _mm_mul_ps(v, isr0);
return mul0; return mul0;
} }
GLM_FUNC_QUALIFIER __m128 glm_f32v4_ffd(__m128 N, __m128 I, __m128 Nref) GLM_FUNC_QUALIFIER __m128 glm_f32v4_ffd(__m128 N, __m128 I, __m128 Nref)
{ {
__m128 dot0 = glm_f32v4_dot(Nref, I); __m128 const dot0 = glm_f32v4_dot(Nref, I);
__m128 sgn0 = glm_f32v4_sgn(dot0); __m128 const sgn0 = glm_f32v4_sgn(dot0);
__m128 mul0 = _mm_mul_ps(sgn0, _mm_set1_ps(-1.0f)); __m128 const mul0 = _mm_mul_ps(sgn0, _mm_set1_ps(-1.0f));
__m128 mul1 = _mm_mul_ps(N, mul0); __m128 const mul1 = _mm_mul_ps(N, mul0);
return mul1; return mul1;
} }
GLM_FUNC_QUALIFIER __m128 glm_f32v4_rfe(__m128 I, __m128 N) GLM_FUNC_QUALIFIER __m128 glm_f32v4_rfe(__m128 I, __m128 N)
{ {
__m128 dot0 = glm_f32v4_dot(N, I); __m128 const dot0 = glm_f32v4_dot(N, I);
__m128 mul0 = _mm_mul_ps(N, dot0); __m128 const mul0 = _mm_mul_ps(N, dot0);
__m128 mul1 = _mm_mul_ps(mul0, _mm_set1_ps(2.0f)); __m128 const mul1 = _mm_mul_ps(mul0, _mm_set1_ps(2.0f));
__m128 sub0 = _mm_sub_ps(I, mul1); __m128 const sub0 = _mm_sub_ps(I, mul1);
return sub0; return sub0;
} }
GLM_FUNC_QUALIFIER __m128 glm_f32v4_rfa(__m128 I, __m128 N, __m128 eta) GLM_FUNC_QUALIFIER __m128 glm_f32v4_rfa(__m128 I, __m128 N, __m128 eta)
{ {
__m128 dot0 = glm_f32v4_dot(N, I); __m128 const dot0 = glm_f32v4_dot(N, I);
__m128 mul0 = _mm_mul_ps(eta, eta); __m128 const mul0 = _mm_mul_ps(eta, eta);
__m128 mul1 = _mm_mul_ps(dot0, dot0); __m128 const mul1 = _mm_mul_ps(dot0, dot0);
__m128 sub0 = _mm_sub_ps(_mm_set1_ps(1.0f), mul0); __m128 const sub0 = _mm_sub_ps(_mm_set1_ps(1.0f), mul0);
__m128 sub1 = _mm_sub_ps(_mm_set1_ps(1.0f), mul1); __m128 const sub1 = _mm_sub_ps(_mm_set1_ps(1.0f), mul1);
__m128 mul2 = _mm_mul_ps(sub0, sub1); __m128 const mul2 = _mm_mul_ps(sub0, sub1);
if(_mm_movemask_ps(_mm_cmplt_ss(mul2, _mm_set1_ps(0.0f))) == 0) if(_mm_movemask_ps(_mm_cmplt_ss(mul2, _mm_set1_ps(0.0f))) == 0)
return _mm_set1_ps(0.0f); return _mm_set1_ps(0.0f);
__m128 sqt0 = _mm_sqrt_ps(mul2); __m128 const sqt0 = _mm_sqrt_ps(mul2);
__m128 mul3 = _mm_mul_ps(eta, dot0); __m128 const mad0 = glm_f32v4_mad(eta, dot0, sqt0);
__m128 add0 = _mm_add_ps(mul3, sqt0); __m128 const mul4 = _mm_mul_ps(mad0, N);
__m128 mul4 = _mm_mul_ps(add0, N); __m128 const mul5 = _mm_mul_ps(eta, I);
__m128 mul5 = _mm_mul_ps(eta, I); __m128 const sub2 = _mm_sub_ps(mul5, mul4);
__m128 sub2 = _mm_sub_ps(mul5, mul4);
return sub2; return sub2;
} }
#endif//GLM_ARCH & GLM_ARCH_SSE2_FLAG #endif//GLM_ARCH & GLM_ARCH_SSE2_BIT

4
glm/simd/integer.h
View File

@ -3,7 +3,7 @@
#pragma once #pragma once
#if GLM_ARCH & GLM_ARCH_SSE2_FLAG #if GLM_ARCH & GLM_ARCH_SSE2_BIT
GLM_FUNC_QUALIFIER __m128i glm_i128_interleave(__m128i x) GLM_FUNC_QUALIFIER __m128i glm_i128_interleave(__m128i x)
{ {
@ -112,4 +112,4 @@ GLM_FUNC_QUALIFIER __m128i glm_i128_interleave2(__m128i x, __m128i y)
return Reg1; return Reg1;
} }
#endif//GLM_ARCH & GLM_ARCH_SSE2_FLAG #endif//GLM_ARCH & GLM_ARCH_SSE2_BIT

4
glm/simd/matrix.h
View File

@ -5,7 +5,7 @@
#include "geometric.h" #include "geometric.h"
#if GLM_ARCH & GLM_ARCH_SSE2_FLAG #if GLM_ARCH & GLM_ARCH_SSE2_BIT
static const __m128 GLM_VAR_USED _m128_rad_ps = _mm_set_ps1(3.141592653589793238462643383279f / 180.f); static const __m128 GLM_VAR_USED _m128_rad_ps = _mm_set_ps1(3.141592653589793238462643383279f / 180.f);
static const __m128 GLM_VAR_USED _m128_deg_ps = _mm_set_ps1(180.f / 3.141592653589793238462643383279f); static const __m128 GLM_VAR_USED _m128_deg_ps = _mm_set_ps1(180.f / 3.141592653589793238462643383279f);
@ -1029,4 +1029,4 @@ GLM_FUNC_QUALIFIER void glm_f32m4_outer(__m128 const & c, __m128 const & r, __m1
out[3] = _mm_mul_ps(c, _mm_shuffle_ps(r, r, _MM_SHUFFLE(3, 3, 3, 3))); out[3] = _mm_mul_ps(c, _mm_shuffle_ps(r, r, _MM_SHUFFLE(3, 3, 3, 3)));
} }
#endif//GLM_ARCH & GLM_ARCH_SSE2_FLAG #endif//GLM_ARCH & GLM_ARCH_SSE2_BIT

4
glm/simd/packing.h
View File

@ -3,6 +3,6 @@
#pragma once #pragma once
#if GLM_ARCH & GLM_ARCH_SSE2_FLAG #if GLM_ARCH & GLM_ARCH_SSE2_BIT
#endif//GLM_ARCH & GLM_ARCH_SSE2_FLAG #endif//GLM_ARCH & GLM_ARCH_SSE2_BIT

66
glm/simd/platform.h
View File

@ -268,34 +268,34 @@
// User defines: GLM_FORCE_PURE GLM_FORCE_SSE2 GLM_FORCE_SSE3 GLM_FORCE_AVX GLM_FORCE_AVX2 GLM_FORCE_AVX2 // User defines: GLM_FORCE_PURE GLM_FORCE_SSE2 GLM_FORCE_SSE3 GLM_FORCE_AVX GLM_FORCE_AVX2 GLM_FORCE_AVX2
#define GLM_ARCH_X86_FLAG 0x00000001 #define GLM_ARCH_X86_BIT 0x00000001
#define GLM_ARCH_SSE2_FLAG 0x00000002 #define GLM_ARCH_SSE2_BIT 0x00000002
#define GLM_ARCH_SSE3_FLAG 0x00000004 #define GLM_ARCH_SSE3_BIT 0x00000004
#define GLM_ARCH_SSSE3_FLAG 0x00000008 #define GLM_ARCH_SSSE3_BIT 0x00000008
#define GLM_ARCH_SSE41_FLAG 0x00000010 #define GLM_ARCH_SSE41_BIT 0x00000010
#define GLM_ARCH_SSE42_FLAG 0x00000020 #define GLM_ARCH_SSE42_BIT 0x00000020
#define GLM_ARCH_AVX_FLAG 0x00000040 #define GLM_ARCH_AVX_BIT 0x00000040
#define GLM_ARCH_AVX2_FLAG 0x00000080 #define GLM_ARCH_AVX2_BIT 0x00000080
#define GLM_ARCH_AVX512_FLAG 0x00000100 // Skylake subset #define GLM_ARCH_AVX512_BIT 0x00000100 // Skylake subset
#define GLM_ARCH_ARM_FLAG 0x00000100 #define GLM_ARCH_ARM_BIT 0x00000100
#define GLM_ARCH_NEON_FLAG 0x00000200 #define GLM_ARCH_NEON_BIT 0x00000200
#define GLM_ARCH_MIPS_FLAG 0x00010000 #define GLM_ARCH_MIPS_BIT 0x00010000
#define GLM_ARCH_PPC_FLAG 0x01000000 #define GLM_ARCH_PPC_BIT 0x01000000
#define GLM_ARCH_PURE (0x00000000) #define GLM_ARCH_PURE (0x00000000)
#define GLM_ARCH_X86 (GLM_ARCH_X86_FLAG) #define GLM_ARCH_X86 (GLM_ARCH_X86_BIT)
#define GLM_ARCH_SSE2 (GLM_ARCH_SSE2_FLAG | GLM_ARCH_X86) #define GLM_ARCH_SSE2 (GLM_ARCH_SSE2_BIT | GLM_ARCH_X86)
#define GLM_ARCH_SSE3 (GLM_ARCH_SSE3_FLAG | GLM_ARCH_SSE2) #define GLM_ARCH_SSE3 (GLM_ARCH_SSE3_BIT | GLM_ARCH_SSE2)
#define GLM_ARCH_SSSE3 (GLM_ARCH_SSSE3_FLAG | GLM_ARCH_SSE3) #define GLM_ARCH_SSSE3 (GLM_ARCH_SSSE3_BIT | GLM_ARCH_SSE3)
#define GLM_ARCH_SSE41 (GLM_ARCH_SSE41_FLAG | GLM_ARCH_SSSE3) #define GLM_ARCH_SSE41 (GLM_ARCH_SSE41_BIT | GLM_ARCH_SSSE3)
#define GLM_ARCH_SSE42 (GLM_ARCH_SSE42_FLAG | GLM_ARCH_SSE41) #define GLM_ARCH_SSE42 (GLM_ARCH_SSE42_BIT | GLM_ARCH_SSE41)
#define GLM_ARCH_AVX (GLM_ARCH_AVX_FLAG | GLM_ARCH_SSE42) #define GLM_ARCH_AVX (GLM_ARCH_AVX_BIT | GLM_ARCH_SSE42)
#define GLM_ARCH_AVX2 (GLM_ARCH_AVX2_FLAG | GLM_ARCH_AVX) #define GLM_ARCH_AVX2 (GLM_ARCH_AVX2_BIT | GLM_ARCH_AVX)
#define GLM_ARCH_AVX512 (GLM_ARCH_AVX512_FLAG | GLM_ARCH_AVX2) // Skylake subset #define GLM_ARCH_AVX512 (GLM_ARCH_AVX512_BIT | GLM_ARCH_AVX2) // Skylake subset
#define GLM_ARCH_ARM (GLM_ARCH_ARM_FLAG) #define GLM_ARCH_ARM (GLM_ARCH_ARM_BIT)
#define GLM_ARCH_NEON (GLM_ARCH_NEON_FLAG | GLM_ARCH_ARM) #define GLM_ARCH_NEON (GLM_ARCH_NEON_BIT | GLM_ARCH_ARM)
#define GLM_ARCH_MIPS (GLM_ARCH_MIPS_FLAG) #define GLM_ARCH_MIPS (GLM_ARCH_MIPS_BIT)
#define GLM_ARCH_PPC (GLM_ARCH_PPC_FLAG) #define GLM_ARCH_PPC (GLM_ARCH_PPC_BIT)
#if defined(GLM_FORCE_PURE) #if defined(GLM_FORCE_PURE)
# define GLM_ARCH GLM_ARCH_PURE # define GLM_ARCH GLM_ARCH_PURE
@ -383,18 +383,18 @@
# include <intrin.h> # include <intrin.h>
#endif #endif
#if GLM_ARCH & GLM_ARCH_AVX2_FLAG #if GLM_ARCH & GLM_ARCH_AVX2_BIT
# include <immintrin.h> # include <immintrin.h>
#elif GLM_ARCH & GLM_ARCH_AVX_FLAG #elif GLM_ARCH & GLM_ARCH_AVX_BIT
# include <immintrin.h> # include <immintrin.h>
#elif GLM_ARCH & GLM_ARCH_SSE42_FLAG #elif GLM_ARCH & GLM_ARCH_SSE42_BIT
# include <nmmintrin.h> # include <nmmintrin.h>
#elif GLM_ARCH & GLM_ARCH_SSE41_FLAG #elif GLM_ARCH & GLM_ARCH_SSE41_BIT
# include <smmintrin.h> # include <smmintrin.h>
#elif GLM_ARCH & GLM_ARCH_SSSE3_FLAG #elif GLM_ARCH & GLM_ARCH_SSSE3_BIT
# include <tmmintrin.h> # include <tmmintrin.h>
#elif GLM_ARCH & GLM_ARCH_SSE3_FLAG #elif GLM_ARCH & GLM_ARCH_SSE3_BIT
# include <pmmintrin.h> # include <pmmintrin.h>
#elif GLM_ARCH & GLM_ARCH_SSE2_FLAG #elif GLM_ARCH & GLM_ARCH_SSE2_BIT
# include <emmintrin.h> # include <emmintrin.h>
#endif//GLM_ARCH #endif//GLM_ARCH

4
glm/simd/trigonometric.h
View File

@ -3,7 +3,7 @@
#pragma once #pragma once
#if GLM_ARCH & GLM_ARCH_SSE2_FLAG #if GLM_ARCH & GLM_ARCH_SSE2_BIT
#endif//GLM_ARCH & GLM_ARCH_SSE2_FLAG #endif//GLM_ARCH & GLM_ARCH_SSE2_BIT

4
glm/simd/vector_relational.h
View File

@ -3,6 +3,6 @@
#pragma once #pragma once
#if GLM_ARCH & GLM_ARCH_SSE2_FLAG #if GLM_ARCH & GLM_ARCH_SSE2_BIT
#endif//GLM_ARCH & GLM_ARCH_SSE2_FLAG #endif//GLM_ARCH & GLM_ARCH_SSE2_BIT

14
test/core/core_setup_message.cpp
View File

@ -176,23 +176,23 @@ int test_instruction_set()
if(GLM_ARCH == GLM_ARCH_PURE) if(GLM_ARCH == GLM_ARCH_PURE)
std::printf("GLM_ARCH_PURE "); std::printf("GLM_ARCH_PURE ");
if(GLM_ARCH & GLM_ARCH_ARM_FLAG) if(GLM_ARCH & GLM_ARCH_ARM_BIT)
std::printf("ARM "); std::printf("ARM ");
if(GLM_ARCH & GLM_ARCH_NEON_FLAG) if(GLM_ARCH & GLM_ARCH_NEON_BIT)
std::printf("NEON "); std::printf("NEON ");
if(GLM_ARCH & GLM_ARCH_AVX2) if(GLM_ARCH & GLM_ARCH_AVX2)
std::printf("AVX2 "); std::printf("AVX2 ");
if(GLM_ARCH & GLM_ARCH_AVX) if(GLM_ARCH & GLM_ARCH_AVX)
std::printf("AVX "); std::printf("AVX ");
if(GLM_ARCH & GLM_ARCH_SSE42_FLAG) if(GLM_ARCH & GLM_ARCH_SSE42_BIT)
std::printf("SSE4.2 "); std::printf("SSE4.2 ");
if(GLM_ARCH & GLM_ARCH_SSE41_FLAG) if(GLM_ARCH & GLM_ARCH_SSE41_BIT)
std::printf("SSE4.1 "); std::printf("SSE4.1 ");
if(GLM_ARCH & GLM_ARCH_SSSE3_FLAG) if(GLM_ARCH & GLM_ARCH_SSSE3_BIT)
std::printf("SSSE3 "); std::printf("SSSE3 ");
if(GLM_ARCH & GLM_ARCH_SSE3_FLAG) if(GLM_ARCH & GLM_ARCH_SSE3_BIT)
std::printf("SSE3 "); std::printf("SSE3 ");
if(GLM_ARCH & GLM_ARCH_SSE2_FLAG) if(GLM_ARCH & GLM_ARCH_SSE2_BIT)
std::printf("SSE2 "); std::printf("SSE2 ");
std::printf("\n"); std::printf("\n");

8
test/gtc/gtc_bitfield.cpp
View File

@ -505,7 +505,7 @@ namespace bitfieldInterleave
assert(A == C); assert(A == C);
assert(A == D); assert(A == D);
# if GLM_ARCH & GLM_ARCH_SSE2_FLAG # if GLM_ARCH & GLM_ARCH_SSE2_BIT
glm::uint64 E = sseBitfieldInterleave(x, y); glm::uint64 E = sseBitfieldInterleave(x, y);
glm::uint64 F = sseUnalignedBitfieldInterleave(x, y); glm::uint64 F = sseUnalignedBitfieldInterleave(x, y);
assert(A == E); assert(A == E);
@ -515,7 +515,7 @@ namespace bitfieldInterleave
glm::uint64 Result[2]; glm::uint64 Result[2];
_mm_storeu_si128((__m128i*)Result, G); _mm_storeu_si128((__m128i*)Result, G);
assert(A == Result[0]); assert(A == Result[0]);
# endif//GLM_ARCH & GLM_ARCH_SSE2_FLAG # endif//GLM_ARCH & GLM_ARCH_SSE2_BIT
} }
} }
@ -629,7 +629,7 @@ namespace bitfieldInterleave
std::printf("glm::detail::bitfieldInterleave Time %d clocks\n", static_cast<unsigned int>(Time)); std::printf("glm::detail::bitfieldInterleave Time %d clocks\n", static_cast<unsigned int>(Time));
} }
# if(GLM_ARCH & GLM_ARCH_SSE2_FLAG && !(GLM_COMPILER & GLM_COMPILER_GCC)) # if(GLM_ARCH & GLM_ARCH_SSE2_BIT && !(GLM_COMPILER & GLM_COMPILER_GCC))
{ {
// SIMD // SIMD
std::vector<__m128i> SimdData; std::vector<__m128i> SimdData;
@ -648,7 +648,7 @@ namespace bitfieldInterleave
std::printf("_mm_bit_interleave_si128 Time %d clocks\n", static_cast<unsigned int>(Time)); std::printf("_mm_bit_interleave_si128 Time %d clocks\n", static_cast<unsigned int>(Time));
} }
# endif//GLM_ARCH & GLM_ARCH_SSE2_FLAG # endif//GLM_ARCH & GLM_ARCH_SSE2_BIT
return 0; return 0;
} }